Get trending papers in your email inbox once a day!
Get trending papers in your email inbox!
SubscribeVALL-E 2: Neural Codec Language Models are Human Parity Zero-Shot Text to Speech Synthesizers
This paper introduces VALL-E 2, the latest advancement in neural codec language models that marks a milestone in zero-shot text-to-speech synthesis (TTS), achieving human parity for the first time. Based on its predecessor, VALL-E, the new iteration introduces two significant enhancements: Repetition Aware Sampling refines the original nucleus sampling process by accounting for token repetition in the decoding history. It not only stabilizes the decoding but also circumvents the infinite loop issue. Grouped Code Modeling organizes codec codes into groups to effectively shorten the sequence length, which not only boosts inference speed but also addresses the challenges of long sequence modeling. Our experiments on the LibriSpeech and VCTK datasets show that VALL-E 2 surpasses previous systems in speech robustness, naturalness, and speaker similarity. It is the first of its kind to reach human parity on these benchmarks. Moreover, VALL-E 2 consistently synthesizes high-quality speech, even for sentences that are traditionally challenging due to their complexity or repetitive phrases. The advantages of this work could contribute to valuable endeavors, such as generating speech for individuals with aphasia or people with amyotrophic lateral sclerosis. Demos of VALL-E 2 will be posted to https://aka.ms/valle2.
DurIAN-E 2: Duration Informed Attention Network with Adaptive Variational Autoencoder and Adversarial Learning for Expressive Text-to-Speech Synthesis
This paper proposes an improved version of DurIAN-E (DurIAN-E 2), which is also a duration informed attention neural network for expressive and high-fidelity text-to-speech (TTS) synthesis. Similar with the DurIAN-E model, multiple stacked SwishRNN-based Transformer blocks are utilized as linguistic encoders and Style-Adaptive Instance Normalization (SAIN) layers are also exploited into frame-level encoders to improve the modeling ability of expressiveness in the proposed the DurIAN-E 2. Meanwhile, motivated by other TTS models using generative models such as VITS, the proposed DurIAN-E 2 utilizes variational autoencoders (VAEs) augmented with normalizing flows and a BigVGAN waveform generator with adversarial training strategy, which further improve the synthesized speech quality and expressiveness. Both objective test and subjective evaluation results prove that the proposed expressive TTS model DurIAN-E 2 can achieve better performance than several state-of-the-art approaches besides DurIAN-E.
Scaling up GANs for Text-to-Image Synthesis
The recent success of text-to-image synthesis has taken the world by storm and captured the general public's imagination. From a technical standpoint, it also marked a drastic change in the favored architecture to design generative image models. GANs used to be the de facto choice, with techniques like StyleGAN. With DALL-E 2, auto-regressive and diffusion models became the new standard for large-scale generative models overnight. This rapid shift raises a fundamental question: can we scale up GANs to benefit from large datasets like LAION? We find that na\"Ively increasing the capacity of the StyleGAN architecture quickly becomes unstable. We introduce GigaGAN, a new GAN architecture that far exceeds this limit, demonstrating GANs as a viable option for text-to-image synthesis. GigaGAN offers three major advantages. First, it is orders of magnitude faster at inference time, taking only 0.13 seconds to synthesize a 512px image. Second, it can synthesize high-resolution images, for example, 16-megapixel pixels in 3.66 seconds. Finally, GigaGAN supports various latent space editing applications such as latent interpolation, style mixing, and vector arithmetic operations.
HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion
Despite significant advances in large-scale text-to-image models, achieving hyper-realistic human image generation remains a desirable yet unsolved task. Existing models like Stable Diffusion and DALL-E 2 tend to generate human images with incoherent parts or unnatural poses. To tackle these challenges, our key insight is that human image is inherently structural over multiple granularities, from the coarse-level body skeleton to fine-grained spatial geometry. Therefore, capturing such correlations between the explicit appearance and latent structure in one model is essential to generate coherent and natural human images. To this end, we propose a unified framework, HyperHuman, that generates in-the-wild human images of high realism and diverse layouts. Specifically, 1) we first build a large-scale human-centric dataset, named HumanVerse, which consists of 340M images with comprehensive annotations like human pose, depth, and surface normal. 2) Next, we propose a Latent Structural Diffusion Model that simultaneously denoises the depth and surface normal along with the synthesized RGB image. Our model enforces the joint learning of image appearance, spatial relationship, and geometry in a unified network, where each branch in the model complements to each other with both structural awareness and textural richness. 3) Finally, to further boost the visual quality, we propose a Structure-Guided Refiner to compose the predicted conditions for more detailed generation of higher resolution. Extensive experiments demonstrate that our framework yields the state-of-the-art performance, generating hyper-realistic human images under diverse scenarios. Project Page: https://snap-research.github.io/HyperHuman/
Extracting Training Data from Diffusion Models
Image diffusion models such as DALL-E 2, Imagen, and Stable Diffusion have attracted significant attention due to their ability to generate high-quality synthetic images. In this work, we show that diffusion models memorize individual images from their training data and emit them at generation time. With a generate-and-filter pipeline, we extract over a thousand training examples from state-of-the-art models, ranging from photographs of individual people to trademarked company logos. We also train hundreds of diffusion models in various settings to analyze how different modeling and data decisions affect privacy. Overall, our results show that diffusion models are much less private than prior generative models such as GANs, and that mitigating these vulnerabilities may require new advances in privacy-preserving training.
Shifted Diffusion for Text-to-image Generation
We present Corgi, a novel method for text-to-image generation. Corgi is based on our proposed shifted diffusion model, which achieves better image embedding generation from input text. Unlike the baseline diffusion model used in DALL-E 2, our method seamlessly encodes prior knowledge of the pre-trained CLIP model in its diffusion process by designing a new initialization distribution and a new transition step of the diffusion. Compared to the strong DALL-E 2 baseline, our method performs better in generating image embedding from the text in terms of both efficiency and effectiveness, resulting in better text-to-image generation. Extensive large-scale experiments are conducted and evaluated in terms of both quantitative measures and human evaluation, indicating a stronger generation ability of our method compared to existing ones. Furthermore, our model enables semi-supervised and language-free training for text-to-image generation, where only part or none of the images in the training dataset have an associated caption. Trained with only 1.7% of the images being captioned, our semi-supervised model obtains FID results comparable to DALL-E 2 on zero-shot text-to-image generation evaluated on MS-COCO. Corgi also achieves new state-of-the-art results across different datasets on downstream language-free text-to-image generation tasks, outperforming the previous method, Lafite, by a large margin.
DS-Fusion: Artistic Typography via Discriminated and Stylized Diffusion
We introduce a novel method to automatically generate an artistic typography by stylizing one or more letter fonts to visually convey the semantics of an input word, while ensuring that the output remains readable. To address an assortment of challenges with our task at hand including conflicting goals (artistic stylization vs. legibility), lack of ground truth, and immense search space, our approach utilizes large language models to bridge texts and visual images for stylization and build an unsupervised generative model with a diffusion model backbone. Specifically, we employ the denoising generator in Latent Diffusion Model (LDM), with the key addition of a CNN-based discriminator to adapt the input style onto the input text. The discriminator uses rasterized images of a given letter/word font as real samples and output of the denoising generator as fake samples. Our model is coined DS-Fusion for discriminated and stylized diffusion. We showcase the quality and versatility of our method through numerous examples, qualitative and quantitative evaluation, as well as ablation studies. User studies comparing to strong baselines including CLIPDraw and DALL-E 2, as well as artist-crafted typographies, demonstrate strong performance of DS-Fusion.
AiGen-FoodReview: A Multimodal Dataset of Machine-Generated Restaurant Reviews and Images on Social Media
Online reviews in the form of user-generated content (UGC) significantly impact consumer decision-making. However, the pervasive issue of not only human fake content but also machine-generated content challenges UGC's reliability. Recent advances in Large Language Models (LLMs) may pave the way to fabricate indistinguishable fake generated content at a much lower cost. Leveraging OpenAI's GPT-4-Turbo and DALL-E-2 models, we craft AiGen-FoodReview, a multi-modal dataset of 20,144 restaurant review-image pairs divided into authentic and machine-generated. We explore unimodal and multimodal detection models, achieving 99.80% multimodal accuracy with FLAVA. We use attributes from readability and photographic theories to score reviews and images, respectively, demonstrating their utility as hand-crafted features in scalable and interpretable detection models, with comparable performance. The paper contributes by open-sourcing the dataset and releasing fake review detectors, recommending its use in unimodal and multimodal fake review detection tasks, and evaluating linguistic and visual features in synthetic versus authentic data.
Generative AI
The term "generative AI" refers to computational techniques that are capable of generating seemingly new, meaningful content such as text, images, or audio from training data. The widespread diffusion of this technology with examples such as Dall-E 2, GPT-4, and Copilot is currently revolutionizing the way we work and communicate with each other. In this article, we provide a conceptualization of generative AI as an entity in socio-technical systems and provide examples of models, systems, and applications. Based on that, we introduce limitations of current generative AI and provide an agenda for Business & Information Systems Engineering (BISE) research. Different from previous works, we focus on generative AI in the context of information systems, and, to this end, we discuss several opportunities and challenges that are unique to the BISE community and make suggestions for impactful directions for BISE research.
SurrogatePrompt: Bypassing the Safety Filter of Text-To-Image Models via Substitution
Advanced text-to-image models such as DALL-E 2 and Midjourney possess the capacity to generate highly realistic images, raising significant concerns regarding the potential proliferation of unsafe content. This includes adult, violent, or deceptive imagery of political figures. Despite claims of rigorous safety mechanisms implemented in these models to restrict the generation of not-safe-for-work (NSFW) content, we successfully devise and exhibit the first prompt attacks on Midjourney, resulting in the production of abundant photorealistic NSFW images. We reveal the fundamental principles of such prompt attacks and suggest strategically substituting high-risk sections within a suspect prompt to evade closed-source safety measures. Our novel framework, SurrogatePrompt, systematically generates attack prompts, utilizing large language models, image-to-text, and image-to-image modules to automate attack prompt creation at scale. Evaluation results disclose an 88% success rate in bypassing Midjourney's proprietary safety filter with our attack prompts, leading to the generation of counterfeit images depicting political figures in violent scenarios. Both subjective and objective assessments validate that the images generated from our attack prompts present considerable safety hazards.
Investigating Prompt Engineering in Diffusion Models
With the spread of the use of Text2Img diffusion models such as DALL-E 2, Imagen, Mid Journey and Stable Diffusion, one challenge that artists face is selecting the right prompts to achieve the desired artistic output. We present techniques for measuring the effect that specific words and phrases in prompts have, and (in the Appendix) present guidance on the selection of prompts to produce desired effects.
Multimodal Latent Language Modeling with Next-Token Diffusion
Multimodal generative models require a unified approach to handle both discrete data (e.g., text and code) and continuous data (e.g., image, audio, video). In this work, we propose Latent Language Modeling (LatentLM), which seamlessly integrates continuous and discrete data using causal Transformers. Specifically, we employ a variational autoencoder (VAE) to represent continuous data as latent vectors and introduce next-token diffusion for autoregressive generation of these vectors. Additionally, we develop sigma-VAE to address the challenges of variance collapse, which is crucial for autoregressive modeling. Extensive experiments demonstrate the effectiveness of LatentLM across various modalities. In image generation, LatentLM surpasses Diffusion Transformers in both performance and scalability. When integrated into multimodal large language models, LatentLM provides a general-purpose interface that unifies multimodal generation and understanding. Experimental results show that LatentLM achieves favorable performance compared to Transfusion and vector quantized models in the setting of scaling up training tokens. In text-to-speech synthesis, LatentLM outperforms the state-of-the-art VALL-E 2 model in speaker similarity and robustness, while requiring 10x fewer decoding steps. The results establish LatentLM as a highly effective and scalable approach to advance large multimodal models.
ECLIPSE: A Resource-Efficient Text-to-Image Prior for Image Generations
Text-to-image (T2I) diffusion models, notably the unCLIP models (e.g., DALL-E-2), achieve state-of-the-art (SOTA) performance on various compositional T2I benchmarks, at the cost of significant computational resources. The unCLIP stack comprises T2I prior and diffusion image decoder. The T2I prior model alone adds a billion parameters compared to the Latent Diffusion Models, which increases the computational and high-quality data requirements. We introduce ECLIPSE, a novel contrastive learning method that is both parameter and data-efficient. ECLIPSE leverages pre-trained vision-language models (e.g., CLIP) to distill the knowledge into the prior model. We demonstrate that the ECLIPSE trained prior, with only 3.3% of the parameters and trained on a mere 2.8% of the data, surpasses the baseline T2I priors with an average of 71.6% preference score under resource-limited setting. It also attains performance on par with SOTA big models, achieving an average of 63.36% preference score in terms of the ability to follow the text compositions. Extensive experiments on two unCLIP diffusion image decoders, Karlo and Kandinsky, affirm that ECLIPSE priors consistently deliver high performance while significantly reducing resource dependency.
Stable Bias: Analyzing Societal Representations in Diffusion Models
As machine learning-enabled Text-to-Image (TTI) systems are becoming increasingly prevalent and seeing growing adoption as commercial services, characterizing the social biases they exhibit is a necessary first step to lowering their risk of discriminatory outcomes. This evaluation, however, is made more difficult by the synthetic nature of these systems' outputs; since artificial depictions of fictive humans have no inherent gender or ethnicity nor do they belong to socially-constructed groups, we need to look beyond common categorizations of diversity or representation. To address this need, we propose a new method for exploring and quantifying social biases in TTI systems by directly comparing collections of generated images designed to showcase a system's variation across social attributes -- gender and ethnicity -- and target attributes for bias evaluation -- professions and gender-coded adjectives. Our approach allows us to (i) identify specific bias trends through visualization tools, (ii) provide targeted scores to directly compare models in terms of diversity and representation, and (iii) jointly model interdependent social variables to support a multidimensional analysis. We use this approach to analyze over 96,000 images generated by 3 popular TTI systems (DALL-E 2, Stable Diffusion v 1.4 and v 2) and find that all three significantly over-represent the portion of their latent space associated with whiteness and masculinity across target attributes; among the systems studied, DALL-E 2 shows the least diversity, followed by Stable Diffusion v2 then v1.4.
LidarCLIP or: How I Learned to Talk to Point Clouds
Research connecting text and images has recently seen several breakthroughs, with models like CLIP, DALL-E 2, and Stable Diffusion. However, the connection between text and other visual modalities, such as lidar data, has received less attention, prohibited by the lack of text-lidar datasets. In this work, we propose LidarCLIP, a mapping from automotive point clouds to a pre-existing CLIP embedding space. Using image-lidar pairs, we supervise a point cloud encoder with the image CLIP embeddings, effectively relating text and lidar data with the image domain as an intermediary. We show the effectiveness of LidarCLIP by demonstrating that lidar-based retrieval is generally on par with image-based retrieval, but with complementary strengths and weaknesses. By combining image and lidar features, we improve upon both single-modality methods and enable a targeted search for challenging detection scenarios under adverse sensor conditions. We also explore zero-shot classification and show that LidarCLIP outperforms existing attempts to use CLIP for point clouds by a large margin. Finally, we leverage our compatibility with CLIP to explore a range of applications, such as point cloud captioning and lidar-to-image generation, without any additional training. Code and pre-trained models are available at https://github.com/atonderski/lidarclip.
GECCO: Geometrically-Conditioned Point Diffusion Models
Diffusion models generating images conditionally on text, such as Dall-E 2 and Stable Diffusion, have recently made a splash far beyond the computer vision community. Here, we tackle the related problem of generating point clouds, both unconditionally, and conditionally with images. For the latter, we introduce a novel geometrically-motivated conditioning scheme based on projecting sparse image features into the point cloud and attaching them to each individual point, at every step in the denoising process. This approach improves geometric consistency and yields greater fidelity than current methods relying on unstructured, global latent codes. Additionally, we show how to apply recent continuous-time diffusion schemes. Our method performs on par or above the state of art on conditional and unconditional experiments on synthetic data, while being faster, lighter, and delivering tractable likelihoods. We show it can also scale to diverse indoors scenes.
A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPT
Recently, ChatGPT, along with DALL-E-2 and Codex,has been gaining significant attention from society. As a result, many individuals have become interested in related resources and are seeking to uncover the background and secrets behind its impressive performance. In fact, ChatGPT and other Generative AI (GAI) techniques belong to the category of Artificial Intelligence Generated Content (AIGC), which involves the creation of digital content, such as images, music, and natural language, through AI models. The goal of AIGC is to make the content creation process more efficient and accessible, allowing for the production of high-quality content at a faster pace. AIGC is achieved by extracting and understanding intent information from instructions provided by human, and generating the content according to its knowledge and the intent information. In recent years, large-scale models have become increasingly important in AIGC as they provide better intent extraction and thus, improved generation results. With the growth of data and the size of the models, the distribution that the model can learn becomes more comprehensive and closer to reality, leading to more realistic and high-quality content generation. This survey provides a comprehensive review on the history of generative models, and basic components, recent advances in AIGC from unimodal interaction and multimodal interaction. From the perspective of unimodality, we introduce the generation tasks and relative models of text and image. From the perspective of multimodality, we introduce the cross-application between the modalities mentioned above. Finally, we discuss the existing open problems and future challenges in AIGC.
Muse: Text-To-Image Generation via Masked Generative Transformers
We present Muse, a text-to-image Transformer model that achieves state-of-the-art image generation performance while being significantly more efficient than diffusion or autoregressive models. Muse is trained on a masked modeling task in discrete token space: given the text embedding extracted from a pre-trained large language model (LLM), Muse is trained to predict randomly masked image tokens. Compared to pixel-space diffusion models, such as Imagen and DALL-E 2, Muse is significantly more efficient due to the use of discrete tokens and requiring fewer sampling iterations; compared to autoregressive models, such as Parti, Muse is more efficient due to the use of parallel decoding. The use of a pre-trained LLM enables fine-grained language understanding, translating to high-fidelity image generation and the understanding of visual concepts such as objects, their spatial relationships, pose, cardinality etc. Our 900M parameter model achieves a new SOTA on CC3M, with an FID score of 6.06. The Muse 3B parameter model achieves an FID of 7.88 on zero-shot COCO evaluation, along with a CLIP score of 0.32. Muse also directly enables a number of image editing applications without the need to fine-tune or invert the model: inpainting, outpainting, and mask-free editing. More results are available at https://muse-model.github.io
STEVE-1: A Generative Model for Text-to-Behavior in Minecraft
Constructing AI models that respond to text instructions is challenging, especially for sequential decision-making tasks. This work introduces an instruction-tuned Video Pretraining (VPT) model for Minecraft called STEVE-1, demonstrating that the unCLIP approach, utilized in DALL-E 2, is also effective for creating instruction-following sequential decision-making agents. STEVE-1 is trained in two steps: adapting the pretrained VPT model to follow commands in MineCLIP's latent space, then training a prior to predict latent codes from text. This allows us to finetune VPT through self-supervised behavioral cloning and hindsight relabeling, bypassing the need for costly human text annotations. By leveraging pretrained models like VPT and MineCLIP and employing best practices from text-conditioned image generation, STEVE-1 costs just $60 to train and can follow a wide range of short-horizon open-ended text and visual instructions in Minecraft. STEVE-1 sets a new bar for open-ended instruction following in Minecraft with low-level controls (mouse and keyboard) and raw pixel inputs, far outperforming previous baselines. We provide experimental evidence highlighting key factors for downstream performance, including pretraining, classifier-free guidance, and data scaling. All resources, including our model weights, training scripts, and evaluation tools are made available for further research.
RAPHAEL: Text-to-Image Generation via Large Mixture of Diffusion Paths
Text-to-image generation has recently witnessed remarkable achievements. We introduce a text-conditional image diffusion model, termed RAPHAEL, to generate highly artistic images, which accurately portray the text prompts, encompassing multiple nouns, adjectives, and verbs. This is achieved by stacking tens of mixture-of-experts (MoEs) layers, i.e., space-MoE and time-MoE layers, enabling billions of diffusion paths (routes) from the network input to the output. Each path intuitively functions as a "painter" for depicting a particular textual concept onto a specified image region at a diffusion timestep. Comprehensive experiments reveal that RAPHAEL outperforms recent cutting-edge models, such as Stable Diffusion, ERNIE-ViLG 2.0, DeepFloyd, and DALL-E 2, in terms of both image quality and aesthetic appeal. Firstly, RAPHAEL exhibits superior performance in switching images across diverse styles, such as Japanese comics, realism, cyberpunk, and ink illustration. Secondly, a single model with three billion parameters, trained on 1,000 A100 GPUs for two months, achieves a state-of-the-art zero-shot FID score of 6.61 on the COCO dataset. Furthermore, RAPHAEL significantly surpasses its counterparts in human evaluation on the ViLG-300 benchmark. We believe that RAPHAEL holds the potential to propel the frontiers of image generation research in both academia and industry, paving the way for future breakthroughs in this rapidly evolving field. More details can be found on a project webpage: https://raphael-painter.github.io/.
Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding
We present Imagen, a text-to-image diffusion model with an unprecedented degree of photorealism and a deep level of language understanding. Imagen builds on the power of large transformer language models in understanding text and hinges on the strength of diffusion models in high-fidelity image generation. Our key discovery is that generic large language models (e.g. T5), pretrained on text-only corpora, are surprisingly effective at encoding text for image synthesis: increasing the size of the language model in Imagen boosts both sample fidelity and image-text alignment much more than increasing the size of the image diffusion model. Imagen achieves a new state-of-the-art FID score of 7.27 on the COCO dataset, without ever training on COCO, and human raters find Imagen samples to be on par with the COCO data itself in image-text alignment. To assess text-to-image models in greater depth, we introduce DrawBench, a comprehensive and challenging benchmark for text-to-image models. With DrawBench, we compare Imagen with recent methods including VQ-GAN+CLIP, Latent Diffusion Models, and DALL-E 2, and find that human raters prefer Imagen over other models in side-by-side comparisons, both in terms of sample quality and image-text alignment. See https://imagen.research.google/ for an overview of the results.
Imagen Editor and EditBench: Advancing and Evaluating Text-Guided Image Inpainting
Text-guided image editing can have a transformative impact in supporting creative applications. A key challenge is to generate edits that are faithful to input text prompts, while consistent with input images. We present Imagen Editor, a cascaded diffusion model built, by fine-tuning Imagen on text-guided image inpainting. Imagen Editor's edits are faithful to the text prompts, which is accomplished by using object detectors to propose inpainting masks during training. In addition, Imagen Editor captures fine details in the input image by conditioning the cascaded pipeline on the original high resolution image. To improve qualitative and quantitative evaluation, we introduce EditBench, a systematic benchmark for text-guided image inpainting. EditBench evaluates inpainting edits on natural and generated images exploring objects, attributes, and scenes. Through extensive human evaluation on EditBench, we find that object-masking during training leads to across-the-board improvements in text-image alignment -- such that Imagen Editor is preferred over DALL-E 2 and Stable Diffusion -- and, as a cohort, these models are better at object-rendering than text-rendering, and handle material/color/size attributes better than count/shape attributes.
CogView2: Faster and Better Text-to-Image Generation via Hierarchical Transformers
The development of the transformer-based text-to-image models are impeded by its slow generation and complexity for high-resolution images. In this work, we put forward a solution based on hierarchical transformers and local parallel auto-regressive generation. We pretrain a 6B-parameter transformer with a simple and flexible self-supervised task, Cross-modal general language model (CogLM), and finetune it for fast super-resolution. The new text-to-image system, CogView2, shows very competitive generation compared to concurrent state-of-the-art DALL-E-2, and naturally supports interactive text-guided editing on images.
Intensional Inheritance Between Concepts: An Information-Theoretic Interpretation
This paper addresses the problem of formalizing and quantifying the concept of "intensional inheritance" between two concepts. We begin by conceiving the intensional inheritance of W from F as the amount of information the proposition "x is F " provides about the proposition "x is W. To flesh this out, we consider concepts F and W defined by sets of properties left{F_{1}, F_{2}, ldots, F_{n}right} and left{W_{1}, W_{2}, ldots, W_{m}right} with associated degrees left{d_{1}, d_{2}, ldots, d_{n}right} and left{e_{1}, e_{2}, ldots, e_{m}right}, respectively, where the properties may overlap. We then derive formulas for the intensional inheritance using both Shannon information theory and algorithmic information theory, incorporating interaction information among properties. We examine a special case where all properties are mutually exclusive and calculate the intensional inheritance in this case in both frameworks. We also derive expressions for P(W mid F) based on the mutual information formula. Finally we consider the relationship between intensional inheritance and conventional set-theoretic "extensional" inheritance, concluding that in our information-theoretic framework, extensional inheritance emerges as a special case of intensional inheritance.
One Transformer Fits All Distributions in Multi-Modal Diffusion at Scale
This paper proposes a unified diffusion framework (dubbed UniDiffuser) to fit all distributions relevant to a set of multi-modal data in one model. Our key insight is -- learning diffusion models for marginal, conditional, and joint distributions can be unified as predicting the noise in the perturbed data, where the perturbation levels (i.e. timesteps) can be different for different modalities. Inspired by the unified view, UniDiffuser learns all distributions simultaneously with a minimal modification to the original diffusion model -- perturbs data in all modalities instead of a single modality, inputs individual timesteps in different modalities, and predicts the noise of all modalities instead of a single modality. UniDiffuser is parameterized by a transformer for diffusion models to handle input types of different modalities. Implemented on large-scale paired image-text data, UniDiffuser is able to perform image, text, text-to-image, image-to-text, and image-text pair generation by setting proper timesteps without additional overhead. In particular, UniDiffuser is able to produce perceptually realistic samples in all tasks and its quantitative results (e.g., the FID and CLIP score) are not only superior to existing general-purpose models but also comparable to the bespoken models (e.g., Stable Diffusion and DALL-E 2) in representative tasks (e.g., text-to-image generation).
Autoregressive Speech Synthesis with Next-Distribution Prediction
We introduce KALL-E, a novel autoregressive (AR) language modeling approach with next-distribution prediction for text-to-speech (TTS) synthesis. Unlike existing methods, KALL-E directly models and predicts the continuous speech distribution conditioned on text without relying on VAE- or diffusion-based components. Specifically, we use WaveVAE to extract continuous speech distributions from waveforms instead of using discrete speech tokens. A single AR language model predicts these continuous speech distributions from text, with a Kullback-Leibler divergence loss as the constraint. Experimental results show that KALL-E outperforms open-source implementations of YourTTS, VALL-E, NaturalSpeech 2, and CosyVoice in terms of naturalness and speaker similarity in zero-shot TTS scenarios. Moreover, KALL-E demonstrates exceptional zero-shot capabilities in emotion and accent cloning. Importantly, KALL-E presents a more straightforward and effective paradigm for using continuous speech representations in TTS. Audio samples are available at: https://zxf-icpc.github.io/kalle/.
On two problems about isogenies of elliptic curves over finite fields
Isogenies occur throughout the theory of elliptic curves. Recently, the cryptographic protocols based on isogenies are considered as candidates of quantum-resistant cryptographic protocols. Given two elliptic curves E_1, E_2 defined over a finite field k with the same trace, there is a nonconstant isogeny beta from E_2 to E_1 defined over k. This study gives out the index of Hom_{it k}(it E_{rm 1},E_{rm 2})beta as a left ideal in End_{it k}(it E_{rm 2}) and figures out the correspondence between isogenies and kernel ideals. In addition, some results about the non-trivial minimal degree of isogenies between the two elliptic curves are also provided.
WenetSpeech4TTS: A 12,800-hour Mandarin TTS Corpus for Large Speech Generation Model Benchmark
With the development of large text-to-speech (TTS) models and scale-up of the training data, state-of-the-art TTS systems have achieved impressive performance. In this paper, we present WenetSpeech4TTS, a multi-domain Mandarin corpus derived from the open-sourced WenetSpeech dataset. Tailored for the text-to-speech tasks, we refined WenetSpeech by adjusting segment boundaries, enhancing the audio quality, and eliminating speaker mixing within each segment. Following a more accurate transcription process and quality-based data filtering process, the obtained WenetSpeech4TTS corpus contains 12,800 hours of paired audio-text data. Furthermore, we have created subsets of varying sizes, categorized by segment quality scores to allow for TTS model training and fine-tuning. VALL-E and NaturalSpeech 2 systems are trained and fine-tuned on these subsets to validate the usability of WenetSpeech4TTS, establishing baselines on benchmark for fair comparison of TTS systems. The corpus and corresponding benchmarks are publicly available on huggingface.
Pattern and Origin for the Extreme $γ$-ray Flares of 3C 454.3 and 3C 279: An Astrophysical Critical Damper?
We apply a Gaussian process method to the extreme gamma-ray flares of 3C 454.3 and 3C 279 to discover the variable patterns and then to investigate the physical origins of the giant flares. The kernels of stochastically driven damped simple harmonic oscillator (SHO), the damped random-walk (DRW), and Matrm ern-3/2 are respectively used to describe the adaptive-binning gamma-ray light curves of the two flares. Our findings show that both the extreme gamma-ray flares of 3C 454.3 and 3C 279 clearly prefer the SHO kernel in the over-damped mode and the Matrm ern-3/2 kernel over the DRW kernel. The resulted SHO and Matrm ern-3/2 power spectral densities (PSDs) are the same for each object, with the index changing from -4 at high frequencies to 0 at low frequencies. The patterns of the two flares are both approaching the critical damping mode with the quality factor Q approx 0.4 (i.e., the damping ratio eta approx 1.25), but with slightly different damping timescales. The characteristic timescale (corresponding to the broken frequency in the PSD) for 3C 454.3 is 2-3 days and 3-5 days for 3C 279. The variable patterns found here suggest that once the system responds to the energy injection disturbance, the release of the energy in the system is finished abruptly. The obtained timescale provides a constraint on the size of energy dissipation region for each source.
BioCLIP 2: Emergent Properties from Scaling Hierarchical Contrastive Learning
Foundation models trained at scale exhibit remarkable emergent behaviors, learning new capabilities beyond their initial training objectives. We find such emergent behaviors in biological vision models via large-scale contrastive vision-language training. To achieve this, we first curate TreeOfLife-200M, comprising 214 million images of living organisms, the largest and most diverse biological organism image dataset to date. We then train BioCLIP 2 on TreeOfLife-200M to distinguish different species. Despite the narrow training objective, BioCLIP 2 yields extraordinary accuracy when applied to various biological visual tasks such as habitat classification and trait prediction. We identify emergent properties in the learned embedding space of BioCLIP 2. At the inter-species level, the embedding distribution of different species aligns closely with functional and ecological meanings (e.g., beak sizes and habitats). At the intra-species level, instead of being diminished, the intra-species variations (e.g., life stages and sexes) are preserved and better separated in subspaces orthogonal to inter-species distinctions. We provide formal proof and analyses to explain why hierarchical supervision and contrastive objectives encourage these emergent properties. Crucially, our results reveal that these properties become increasingly significant with larger-scale training data, leading to a biologically meaningful embedding space.
An X-ray Significantly Variable, Luminous, Type 2 Quasar at z = 2.99 with a Massive Host Galaxy
We present a comprehensive X-ray analysis and spectral energy distribution (SED) fitting of WISEA J171419.96+602724.6, an extremely luminous type 2 quasar at z = 2.99. The source was suggested as a candidate Compton-thick (column density N_{rm H}>1.5 times 10^{24} cm^{-2}) quasar by a short XMM-Newton observation in 2011. We recently observed the source with deep NuSTAR and XMM-Newton exposures in 2021 and found that the source has a lower obscuration of N_{rm H}sim5 times 10^{22} cm^{-2} with an about four times lower flux. The two epochs of observations suggested that the source was significantly variable in X-ray obscuration, flux, and intrinsic luminosity at 2-3~sigma in less than 2.5 years (in the source rest frame). We performed SED fitting of this source using CIGALE thanks to its great availability of multiwavelength data (from hard X-rays to radio). The source is very luminous with a bolometric luminosity of L_{rm BOL}sim 2.5 times 10^{47} erg s^{-1}. Its host galaxy has a huge star formation rate (SFR) of sim1280 Solar mass yr^{-1} and a huge stellar mass of sim1.1 times 10^{12} Solar mass. The correlation between the SFR and stellar mass of this source is consistent with what was measured in the high-z quasars. It is also consistent with what was measured in the main-sequence star-forming galaxies, suggesting that the presence of the active nucleus in our target does not enhance or suppress the SFR of its host galaxy. The source is an Infrared hyper-luminous, obscured galaxy with significant amount of hot dust in its torus and shares many similar properties with hot, dust obscured galaxies.
The Seventeenth Data Release of the Sloan Digital Sky Surveys: Complete Release of MaNGA, MaStar and APOGEE-2 Data
This paper documents the seventeenth data release (DR17) from the Sloan Digital Sky Surveys; the fifth and final release from the fourth phase (SDSS-IV). DR17 contains the complete release of the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey, which reached its goal of surveying over 10,000 nearby galaxies. The complete release of the MaNGA Stellar Library (MaStar) accompanies this data, providing observations of almost 30,000 stars through the MaNGA instrument during bright time. DR17 also contains the complete release of the Apache Point Observatory Galactic Evolution Experiment 2 (APOGEE-2) survey which publicly releases infra-red spectra of over 650,000 stars. The main sample from the Extended Baryon Oscillation Spectroscopic Survey (eBOSS), as well as the sub-survey Time Domain Spectroscopic Survey (TDSS) data were fully released in DR16. New single-fiber optical spectroscopy released in DR17 is from the SPectroscipic IDentification of ERosita Survey (SPIDERS) sub-survey and the eBOSS-RM program. Along with the primary data sets, DR17 includes 25 new or updated Value Added Catalogs (VACs). This paper concludes the release of SDSS-IV survey data. SDSS continues into its fifth phase with observations already underway for the Milky Way Mapper (MWM), Local Volume Mapper (LVM) and Black Hole Mapper (BHM) surveys.
EC-Guide: A Comprehensive E-Commerce Guide for Instruction Tuning and Quantization
Large language models (LLMs) have attracted considerable attention in various fields for their cost-effective solutions to diverse challenges, especially with advancements in instruction tuning and quantization. E-commerce, with its complex tasks and extensive product-user interactions, presents a promising application area for LLMs. However, the domain-specific concepts and knowledge inherent in e-commerce pose significant challenges for adapting general LLMs. To address this issue, we developed EC-Guide https://github.com/fzp0424/EC-Guide-KDDUP-2024, a comprehensive e-commerce guide for instruction tuning and quantization of LLMs. We also heuristically integrated Chain-of-Thought (CoT) during inference to enhance arithmetic performance. Our approach achieved the 2nd place in Track 2 and 5th place in Track 5 at the Amazon KDD Cup'24 https://www.aicrowd.com/challenges/amazon-kdd-cup-2024-multi-task-online-shopping-challenge-for-llms. Additionally, our solution is model-agnostic, enabling effective scalability across larger systems.
Point-E: A System for Generating 3D Point Clouds from Complex Prompts
While recent work on text-conditional 3D object generation has shown promising results, the state-of-the-art methods typically require multiple GPU-hours to produce a single sample. This is in stark contrast to state-of-the-art generative image models, which produce samples in a number of seconds or minutes. In this paper, we explore an alternative method for 3D object generation which produces 3D models in only 1-2 minutes on a single GPU. Our method first generates a single synthetic view using a text-to-image diffusion model, and then produces a 3D point cloud using a second diffusion model which conditions on the generated image. While our method still falls short of the state-of-the-art in terms of sample quality, it is one to two orders of magnitude faster to sample from, offering a practical trade-off for some use cases. We release our pre-trained point cloud diffusion models, as well as evaluation code and models, at https://github.com/openai/point-e.
EVADE: Multimodal Benchmark for Evasive Content Detection in E-Commerce Applications
E-commerce platforms increasingly rely on Large Language Models (LLMs) and Vision-Language Models (VLMs) to detect illicit or misleading product content. However, these models remain vulnerable to evasive content: inputs (text or images) that superficially comply with platform policies while covertly conveying prohibited claims. Unlike traditional adversarial attacks that induce overt failures, evasive content exploits ambiguity and context, making it far harder to detect. Existing robustness benchmarks provide little guidance for this demanding, real-world challenge. We introduce EVADE, the first expert-curated, Chinese, multimodal benchmark specifically designed to evaluate foundation models on evasive content detection in e-commerce. The dataset contains 2,833 annotated text samples and 13,961 images spanning six demanding product categories, including body shaping, height growth, and health supplements. Two complementary tasks assess distinct capabilities: Single-Violation, which probes fine-grained reasoning under short prompts, and All-in-One, which tests long-context reasoning by merging overlapping policy rules into unified instructions. Notably, the All-in-One setting significantly narrows the performance gap between partial and full-match accuracy, suggesting that clearer rule definitions improve alignment between human and model judgment. We benchmark 26 mainstream LLMs and VLMs and observe substantial performance gaps: even state-of-the-art models frequently misclassify evasive samples. By releasing EVADE and strong baselines, we provide the first rigorous standard for evaluating evasive-content detection, expose fundamental limitations in current multimodal reasoning, and lay the groundwork for safer and more transparent content moderation systems in e-commerce. The dataset is publicly available at https://huggingface.co/datasets/koenshen/EVADE-Bench.
E-MD3C: Taming Masked Diffusion Transformers for Efficient Zero-Shot Object Customization
We propose E-MD3C (Efficient Masked Diffusion Transformer with Disentangled Conditions and Compact Collector), a highly efficient framework for zero-shot object image customization. Unlike prior works reliant on resource-intensive Unet architectures, our approach employs lightweight masked diffusion transformers operating on latent patches, offering significantly improved computational efficiency. The framework integrates three core components: (1) an efficient masked diffusion transformer for processing autoencoder latents, (2) a disentangled condition design that ensures compactness while preserving background alignment and fine details, and (3) a learnable Conditions Collector that consolidates multiple inputs into a compact representation for efficient denoising and learning. E-MD3C outperforms the existing approach on the VITON-HD dataset across metrics such as PSNR, FID, SSIM, and LPIPS, demonstrating clear advantages in parameters, memory efficiency, and inference speed. With only 1{4} of the parameters, our Transformer-based 468M model delivers 2.5times faster inference and uses 2{3} of the GPU memory compared to an 1720M Unet-based latent diffusion model.
E-CAR: Efficient Continuous Autoregressive Image Generation via Multistage Modeling
Recent advances in autoregressive (AR) models with continuous tokens for image generation show promising results by eliminating the need for discrete tokenization. However, these models face efficiency challenges due to their sequential token generation nature and reliance on computationally intensive diffusion-based sampling. We present ECAR (Efficient Continuous Auto-Regressive Image Generation via Multistage Modeling), an approach that addresses these limitations through two intertwined innovations: (1) a stage-wise continuous token generation strategy that reduces computational complexity and provides progressively refined token maps as hierarchical conditions, and (2) a multistage flow-based distribution modeling method that transforms only partial-denoised distributions at each stage comparing to complete denoising in normal diffusion models. Holistically, ECAR operates by generating tokens at increasing resolutions while simultaneously denoising the image at each stage. This design not only reduces token-to-image transformation cost by a factor of the stage number but also enables parallel processing at the token level. Our approach not only enhances computational efficiency but also aligns naturally with image generation principles by operating in continuous token space and following a hierarchical generation process from coarse to fine details. Experimental results demonstrate that ECAR achieves comparable image quality to DiT Peebles & Xie [2023] while requiring 10times FLOPs reduction and 5times speedup to generate a 256times256 image.
SimpleSpeech 2: Towards Simple and Efficient Text-to-Speech with Flow-based Scalar Latent Transformer Diffusion Models
Scaling Text-to-speech (TTS) to large-scale datasets has been demonstrated as an effective method for improving the diversity and naturalness of synthesized speech. At the high level, previous large-scale TTS models can be categorized into either Auto-regressive (AR) based (e.g., VALL-E) or Non-auto-regressive (NAR) based models (e.g., NaturalSpeech 2/3). Although these works demonstrate good performance, they still have potential weaknesses. For instance, AR-based models are plagued by unstable generation quality and slow generation speed; meanwhile, some NAR-based models need phoneme-level duration alignment information, thereby increasing the complexity of data pre-processing, model design, and loss design. In this work, we build upon our previous publication by implementing a simple and efficient non-autoregressive (NAR) TTS framework, termed SimpleSpeech 2. SimpleSpeech 2 effectively combines the strengths of both autoregressive (AR) and non-autoregressive (NAR) methods, offering the following key advantages: (1) simplified data preparation; (2) straightforward model and loss design; and (3) stable, high-quality generation performance with fast inference speed. Compared to our previous publication, we present ({\romannumeral1}) a detailed analysis of the influence of speech tokenizer and noisy label for TTS performance; ({\romannumeral2}) four distinct types of sentence duration predictors; ({\romannumeral3}) a novel flow-based scalar latent transformer diffusion model. With these improvement, we show a significant improvement in generation performance and generation speed compared to our previous work and other state-of-the-art (SOTA) large-scale TTS models. Furthermore, we show that SimpleSpeech 2 can be seamlessly extended to multilingual TTS by training it on multilingual speech datasets. Demos are available on: {https://dongchaoyang.top/SimpleSpeech2\_demo/}.
ChineseEcomQA: A Scalable E-commerce Concept Evaluation Benchmark for Large Language Models
With the increasing use of Large Language Models (LLMs) in fields such as e-commerce, domain-specific concept evaluation benchmarks are crucial for assessing their domain capabilities. Existing LLMs may generate factually incorrect information within the complex e-commerce applications. Therefore, it is necessary to build an e-commerce concept benchmark. Existing benchmarks encounter two primary challenges: (1) handle the heterogeneous and diverse nature of tasks, (2) distinguish between generality and specificity within the e-commerce field. To address these problems, we propose ChineseEcomQA, a scalable question-answering benchmark focused on fundamental e-commerce concepts. ChineseEcomQA is built on three core characteristics: Focus on Fundamental Concept, E-commerce Generality and E-commerce Expertise. Fundamental concepts are designed to be applicable across a diverse array of e-commerce tasks, thus addressing the challenge of heterogeneity and diversity. Additionally, by carefully balancing generality and specificity, ChineseEcomQA effectively differentiates between broad e-commerce concepts, allowing for precise validation of domain capabilities. We achieve this through a scalable benchmark construction process that combines LLM validation, Retrieval-Augmented Generation (RAG) validation, and rigorous manual annotation. Based on ChineseEcomQA, we conduct extensive evaluations on mainstream LLMs and provide some valuable insights. We hope that ChineseEcomQA could guide future domain-specific evaluations, and facilitate broader LLM adoption in e-commerce applications.
Finding the Subjective Truth: Collecting 2 Million Votes for Comprehensive Gen-AI Model Evaluation
Efficiently evaluating the performance of text-to-image models is difficult as it inherently requires subjective judgment and human preference, making it hard to compare different models and quantify the state of the art. Leveraging Rapidata's technology, we present an efficient annotation framework that sources human feedback from a diverse, global pool of annotators. Our study collected over 2 million annotations across 4,512 images, evaluating four prominent models (DALL-E 3, Flux.1, MidJourney, and Stable Diffusion) on style preference, coherence, and text-to-image alignment. We demonstrate that our approach makes it feasible to comprehensively rank image generation models based on a vast pool of annotators and show that the diverse annotator demographics reflect the world population, significantly decreasing the risk of biases.
SEED-Bench-2: Benchmarking Multimodal Large Language Models
Multimodal large language models (MLLMs), building upon the foundation of powerful large language models (LLMs), have recently demonstrated exceptional capabilities in generating not only texts but also images given interleaved multimodal inputs (acting like a combination of GPT-4V and DALL-E 3). However, existing MLLM benchmarks remain limited to assessing only models' comprehension ability of single image-text inputs, failing to keep up with the strides made in MLLMs. A comprehensive benchmark is imperative for investigating the progress and uncovering the limitations of current MLLMs. In this work, we categorize the capabilities of MLLMs into hierarchical levels from L_0 to L_4 based on the modalities they can accept and generate, and propose SEED-Bench-2, a comprehensive benchmark that evaluates the hierarchical capabilities of MLLMs. Specifically, SEED-Bench-2 comprises 24K multiple-choice questions with accurate human annotations, which spans 27 dimensions, including the evaluation of both text and image generation. Multiple-choice questions with groundtruth options derived from human annotation enables an objective and efficient assessment of model performance, eliminating the need for human or GPT intervention during evaluation. We further evaluate the performance of 23 prominent open-source MLLMs and summarize valuable observations. By revealing the limitations of existing MLLMs through extensive evaluations, we aim for SEED-Bench-2 to provide insights that will motivate future research towards the goal of General Artificial Intelligence. Dataset and evaluation code are available at https://github.com/AILab-CVC/SEED-Bench
LLaSA: Large Language and E-Commerce Shopping Assistant
The e-commerce platform has evolved rapidly due to its widespread popularity and convenience. Developing an e-commerce shopping assistant for customers is crucial to aiding them in quickly finding desired products and recommending precisely what they need. However, most previous shopping assistants face two main problems: (1) task-specificity, which necessitates the development of different models for various tasks, thereby increasing development costs and limiting effectiveness; and (2) poor generalization, where the trained model performs inadequately on up-to-date products. To resolve these issues, we employ Large Language Models (LLMs) to construct an omnipotent assistant, leveraging their adeptness at handling multiple tasks and their superior generalization capability. Nonetheless, LLMs lack inherent knowledge of e-commerce concepts. To address this, we create an instruction dataset comprising 65,000 samples and diverse tasks, termed as EshopInstruct. Through instruction tuning on our dataset, the assistant, named LLaSA, demonstrates the potential to function as an omnipotent assistant. Additionally, we propose various inference optimization strategies to enhance performance with limited inference resources. In the Amazon KDD Cup 2024 Challenge, our proposed method, LLaSA, achieved an overall ranking of 3rd place on ShopBench, including 57 tasks and approximately 20,000 questions, and we secured top-5 rankings in each track, especially in track4, where we achieved the best performance result among all student teams. Our extensive practices fully demonstrate that LLMs possess the great potential to be competent e-commerce shopping assistants.
Captions Speak Louder than Images (CASLIE): Generalizing Foundation Models for E-commerce from High-quality Multimodal Instruction Data
Leveraging multimodal data to drive breakthroughs in e-commerce applications through Multimodal Foundation Models (MFMs) is gaining increasing attention from the research community. However, there are significant challenges that hinder the optimal use of multimodal e-commerce data by foundation models: (1) the scarcity of large-scale, high-quality multimodal benchmark datasets; and (2) the lack of effective multimodal information integration methods. To address these challenges, in this paper, we introduce MMECInstruct, the first-ever, large-scale, and high-quality multimodal instruction dataset for e-commerce. We also develop CASLIE, a simple, lightweight, yet effective framework for integrating multimodal information for e-commerce. Leveraging MMECInstruct, we fine-tune a series of e-commerce MFMs within CASLIE, denoted as CASLIE models. Our comprehensive evaluation demonstrates that CASLIE models substantially outperform 5 categories of advanced baseline models in the in-domain evaluation. Moreover, CASLIE models show strong generalizability to out-of-domain settings. MMECInstruct and CASLIE models are publicly accessible through https://ninglab.github.io/CASLIE/.
VirtualModel: Generating Object-ID-retentive Human-object Interaction Image by Diffusion Model for E-commerce Marketing
Due to the significant advances in large-scale text-to-image generation by diffusion model (DM), controllable human image generation has been attracting much attention recently. Existing works, such as Controlnet [36], T2I-adapter [20] and HumanSD [10] have demonstrated good abilities in generating human images based on pose conditions, they still fail to meet the requirements of real e-commerce scenarios. These include (1) the interaction between the shown product and human should be considered, (2) human parts like face/hand/arm/foot and the interaction between human model and product should be hyper-realistic, and (3) the identity of the product shown in advertising should be exactly consistent with the product itself. To this end, in this paper, we first define a new human image generation task for e-commerce marketing, i.e., Object-ID-retentive Human-object Interaction image Generation (OHG), and then propose a VirtualModel framework to generate human images for product shown, which supports displays of any categories of products and any types of human-object interaction. As shown in Figure 1, VirtualModel not only outperforms other methods in terms of accurate pose control and image quality but also allows for the display of user-specified product objects by maintaining the product-ID consistency and enhancing the plausibility of human-object interaction. Codes and data will be released.
SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
Language models have outpaced our ability to evaluate them effectively, but for their future development it is essential to study the frontier of their capabilities. We consider real-world software engineering to be a rich, sustainable, and challenging testbed for evaluating the next generation of language models. We therefore introduce SWE-bench, an evaluation framework including 2,294 software engineering problems drawn from real GitHub issues and corresponding pull requests across 12 popular Python repositories. Given a codebase along with a description of an issue to be resolved, a language model is tasked with editing the codebase to address the issue. Resolving issues in SWE-bench frequently requires understanding and coordinating changes across multiple functions, classes, and even files simultaneously, calling for models to interact with execution environments, process extremely long contexts and perform complex reasoning that goes far beyond traditional code generation. Our evaluations show that both state-of-the-art proprietary models and our fine-tuned model SWE-Llama can resolve only the simplest issues. Claude 2 and GPT-4 solve a mere 4.8% and 1.7% of instances respectively, even when provided with an oracle retriever. Advances on SWE-bench represent steps towards LMs that are more practical, intelligent, and autonomous.
ASPCAP: The Apogee Stellar Parameter and Chemical Abundances Pipeline
The Apache Point Observatory Galactic Evolution Experiment (APOGEE) has built the largest moderately high-resolution (R=22, 500) spectroscopic map of the stars across the Milky Way, and including dust-obscured areas. The APOGEE Stellar Parameter and Chemical Abundances Pipeline (ASPCAP) is the software developed for the automated analysis of these spectra. ASPCAP determines atmospheric parameters and chemical abundances from observed spectra by comparing observed spectra to libraries of theoretical spectra, using chi-2 minimization in a multidimensional parameter space. The package consists of a fortran90 code that does the actual minimization, and a wrapper IDL code for book-keeping and data handling. This paper explains in detail the ASPCAP components and functionality, and presents results from a number of tests designed to check its performance. ASPCAP provides stellar effective temperatures, surface gravities, and metallicities precise to 2%, 0.1 dex, and 0.05 dex, respectively, for most APOGEE stars, which are predominantly giants. It also provides abundances for up to 15 chemical elements with various levels of precision, typically under 0.1 dex. The final data release (DR12) of the Sloan Digital Sky Survey III contains an APOGEE database of more than 150,000 stars. ASPCAP development continues in the SDSS-IV APOGEE-2 survey.
BEV-Seg: Bird's Eye View Semantic Segmentation Using Geometry and Semantic Point Cloud
Bird's-eye-view (BEV) is a powerful and widely adopted representation for road scenes that captures surrounding objects and their spatial locations, along with overall context in the scene. In this work, we focus on bird's eye semantic segmentation, a task that predicts pixel-wise semantic segmentation in BEV from side RGB images. This task is made possible by simulators such as Carla, which allow for cheap data collection, arbitrary camera placements, and supervision in ways otherwise not possible in the real world. There are two main challenges to this task: the view transformation from side view to bird's eye view, as well as transfer learning to unseen domains. Existing work transforms between views through fully connected layers and transfer learns via GANs. This suffers from a lack of depth reasoning and performance degradation across domains. Our novel 2-staged perception pipeline explicitly predicts pixel depths and combines them with pixel semantics in an efficient manner, allowing the model to leverage depth information to infer objects' spatial locations in the BEV. In addition, we transfer learning by abstracting high-level geometric features and predicting an intermediate representation that is common across different domains. We publish a new dataset called BEVSEG-Carla and show that our approach improves state-of-the-art by 24% mIoU and performs well when transferred to a new domain.
Benchmarking Neural Network Training Algorithms
Training algorithms, broadly construed, are an essential part of every deep learning pipeline. Training algorithm improvements that speed up training across a wide variety of workloads (e.g., better update rules, tuning protocols, learning rate schedules, or data selection schemes) could save time, save computational resources, and lead to better, more accurate, models. Unfortunately, as a community, we are currently unable to reliably identify training algorithm improvements, or even determine the state-of-the-art training algorithm. In this work, using concrete experiments, we argue that real progress in speeding up training requires new benchmarks that resolve three basic challenges faced by empirical comparisons of training algorithms: (1) how to decide when training is complete and precisely measure training time, (2) how to handle the sensitivity of measurements to exact workload details, and (3) how to fairly compare algorithms that require hyperparameter tuning. In order to address these challenges, we introduce a new, competitive, time-to-result benchmark using multiple workloads running on fixed hardware, the AlgoPerf: Training Algorithms benchmark. Our benchmark includes a set of workload variants that make it possible to detect benchmark submissions that are more robust to workload changes than current widely-used methods. Finally, we evaluate baseline submissions constructed using various optimizers that represent current practice, as well as other optimizers that have recently received attention in the literature. These baseline results collectively demonstrate the feasibility of our benchmark, show that non-trivial gaps between methods exist, and set a provisional state-of-the-art for future benchmark submissions to try and surpass.
On the Societal Impact of Open Foundation Models
Foundation models are powerful technologies: how they are released publicly directly shapes their societal impact. In this position paper, we focus on open foundation models, defined here as those with broadly available model weights (e.g. Llama 2, Stable Diffusion XL). We identify five distinctive properties (e.g. greater customizability, poor monitoring) of open foundation models that lead to both their benefits and risks. Open foundation models present significant benefits, with some caveats, that span innovation, competition, the distribution of decision-making power, and transparency. To understand their risks of misuse, we design a risk assessment framework for analyzing their marginal risk. Across several misuse vectors (e.g. cyberattacks, bioweapons), we find that current research is insufficient to effectively characterize the marginal risk of open foundation models relative to pre-existing technologies. The framework helps explain why the marginal risk is low in some cases, clarifies disagreements about misuse risks by revealing that past work has focused on different subsets of the framework with different assumptions, and articulates a way forward for more constructive debate. Overall, our work helps support a more grounded assessment of the societal impact of open foundation models by outlining what research is needed to empirically validate their theoretical benefits and risks.
Ultra Fast Transformers on FPGAs for Particle Physics Experiments
This work introduces a highly efficient implementation of the transformer architecture on a Field-Programmable Gate Array (FPGA) by using the hls4ml tool. Given the demonstrated effectiveness of transformer models in addressing a wide range of problems, their application in experimental triggers within particle physics becomes a subject of significant interest. In this work, we have implemented critical components of a transformer model, such as multi-head attention and softmax layers. To evaluate the effectiveness of our implementation, we have focused on a particle physics jet flavor tagging problem, employing a public dataset. We recorded latency under 2 mus on the Xilinx UltraScale+ FPGA, which is compatible with hardware trigger requirements at the CERN Large Hadron Collider experiments.
QuantumBind-RBFE: Accurate Relative Binding Free Energy Calculations Using Neural Network Potentials
Accurate prediction of protein-ligand binding affinities is crucial in drug discovery, particularly during hit-to-lead and lead optimization phases, however, limitations in ligand force fields continue to impact prediction accuracy. In this work, we validate relative binding free energy (RBFE) accuracy using neural network potentials (NNPs) for the ligands. We utilize a novel NNP model, AceForce 1.0, based on the TensorNet architecture for small molecules that broadens the applicability to diverse drug-like compounds, including all important chemical elements and supporting charged molecules. Using established benchmarks, we show overall improved accuracy and correlation in binding affinity predictions compared with GAFF2 for molecular mechanics and ANI2-x for NNPs. Slightly less accuracy but comparable correlations with OPLS4. We also show that we can run the NNP simulations at 2 fs timestep, at least two times larger than previous NNP models, providing significant speed gains. The results show promise for further evolutions of free energy calculations using NNPs while demonstrating its practical use already with the current generation. The code and NNP model are publicly available for research use.
Large Legal Fictions: Profiling Legal Hallucinations in Large Language Models
Large language models (LLMs) have the potential to transform the practice of law, but this potential is threatened by the presence of legal hallucinations -- responses from these models that are not consistent with legal facts. We investigate the extent of these hallucinations using an original suite of legal queries, comparing LLMs' responses to structured legal metadata and examining their consistency. Our work makes four key contributions: (1) We develop a typology of legal hallucinations, providing a conceptual framework for future research in this area. (2) We find that legal hallucinations are alarmingly prevalent, occurring between 69% of the time with ChatGPT 3.5 and 88% with Llama 2, when these models are asked specific, verifiable questions about random federal court cases. (3) We illustrate that LLMs often fail to correct a user's incorrect legal assumptions in a contra-factual question setup. (4) We provide evidence that LLMs cannot always predict, or do not always know, when they are producing legal hallucinations. Taken together, these findings caution against the rapid and unsupervised integration of popular LLMs into legal tasks. Even experienced lawyers must remain wary of legal hallucinations, and the risks are highest for those who stand to benefit from LLMs the most -- pro se litigants or those without access to traditional legal resources.
On-Policy Policy Gradient Reinforcement Learning Without On-Policy Sampling
On-policy reinforcement learning (RL) algorithms perform policy updates using i.i.d. trajectories collected by the current policy. However, after observing only a finite number of trajectories, on-policy sampling may produce data that fails to match the expected on-policy data distribution. This sampling error leads to noisy updates and data inefficient on-policy learning. Recent work in the policy evaluation setting has shown that non-i.i.d., off-policy sampling can produce data with lower sampling error than on-policy sampling can produce. Motivated by this observation, we introduce an adaptive, off-policy sampling method to improve the data efficiency of on-policy policy gradient algorithms. Our method, Proximal Robust On-Policy Sampling (PROPS), reduces sampling error by collecting data with a behavior policy that increases the probability of sampling actions that are under-sampled with respect to the current policy. Rather than discarding data from old policies -- as is commonly done in on-policy algorithms -- PROPS uses data collection to adjust the distribution of previously collected data to be approximately on-policy. We empirically evaluate PROPS on both continuous-action MuJoCo benchmark tasks as well as discrete-action tasks and demonstrate that (1) PROPS decreases sampling error throughout training and (2) improves the data efficiency of on-policy policy gradient algorithms. Our work improves the RL community's understanding of a nuance in the on-policy vs off-policy dichotomy: on-policy learning requires on-policy data, not on-policy sampling.
Understanding when Dynamics-Invariant Data Augmentations Benefit Model-Free Reinforcement Learning Updates
Recently, data augmentation (DA) has emerged as a method for leveraging domain knowledge to inexpensively generate additional data in reinforcement learning (RL) tasks, often yielding substantial improvements in data efficiency. While prior work has demonstrated the utility of incorporating augmented data directly into model-free RL updates, it is not well-understood when a particular DA strategy will improve data efficiency. In this paper, we seek to identify general aspects of DA responsible for observed learning improvements. Our study focuses on sparse-reward tasks with dynamics-invariant data augmentation functions, serving as an initial step towards a more general understanding of DA and its integration into RL training. Experimentally, we isolate three relevant aspects of DA: state-action coverage, reward density, and the number of augmented transitions generated per update (the augmented replay ratio). From our experiments, we draw two conclusions: (1) increasing state-action coverage often has a much greater impact on data efficiency than increasing reward density, and (2) decreasing the augmented replay ratio substantially improves data efficiency. In fact, certain tasks in our empirical study are solvable only when the replay ratio is sufficiently low.
Tailor: Generating and Perturbing Text with Semantic Controls
Controlled text perturbation is useful for evaluating and improving model generalizability. However, current techniques rely on training a model for every target perturbation, which is expensive and hard to generalize. We present Tailor, a semantically-controlled text generation system. Tailor builds on a pretrained seq2seq model and produces textual outputs conditioned on control codes derived from semantic representations. We craft a set of operations to modify the control codes, which in turn steer generation towards targeted attributes. These operations can be further composed into higher-level ones, allowing for flexible perturbation strategies. We demonstrate the effectiveness of these perturbations in multiple applications. First, we use Tailor to automatically create high-quality contrast sets for four distinct natural language processing (NLP) tasks. These contrast sets contain fewer spurious artifacts and are complementary to manually annotated ones in their lexical diversity. Second, we show that Tailor perturbations can improve model generalization through data augmentation. Perturbing just 2% of training data leads to a 5.8-point gain on an NLI challenge set measuring reliance on syntactic heuristics.
Authorship Attribution in Bangla literature using Character-level CNN
Characters are the smallest unit of text that can extract stylometric signals to determine the author of a text. In this paper, we investigate the effectiveness of character-level signals in Authorship Attribution of Bangla Literature and show that the results are promising but improvable. The time and memory efficiency of the proposed model is much higher than the word level counterparts but accuracy is 2-5% less than the best performing word-level models. Comparison of various word-based models is performed and shown that the proposed model performs increasingly better with larger datasets. We also analyze the effect of pre-training character embedding of diverse Bangla character set in authorship attribution. It is seen that the performance is improved by up to 10% on pre-training. We used 2 datasets from 6 to 14 authors, balancing them before training and compare the results.
Deep contextualized word representations
We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.
Harnessing Business and Media Insights with Large Language Models
This paper introduces Fortune Analytics Language Model (FALM). FALM empowers users with direct access to comprehensive business analysis, including market trends, company performance metrics, and expert insights. Unlike generic LLMs, FALM leverages a curated knowledge base built from professional journalism, enabling it to deliver precise and in-depth answers to intricate business questions. Users can further leverage natural language queries to directly visualize financial data, generating insightful charts and graphs to understand trends across diverse business sectors clearly. FALM fosters user trust and ensures output accuracy through three novel methods: 1) Time-aware reasoning guarantees accurate event registration and prioritizes recent updates. 2) Thematic trend analysis explicitly examines topic evolution over time, providing insights into emerging business landscapes. 3) Content referencing and task decomposition enhance answer fidelity and data visualization accuracy. We conduct both automated and human evaluations, demonstrating FALM's significant performance improvements over baseline methods while prioritizing responsible AI practices. These benchmarks establish FALM as a cutting-edge LLM in the business and media domains, with exceptional accuracy and trustworthiness.
An Empirical Study on Developers Shared Conversations with ChatGPT in GitHub Pull Requests and Issues
ChatGPT has significantly impacted software development practices, providing substantial assistance to developers in a variety of tasks, including coding, testing, and debugging. Despite its widespread adoption, the impact of ChatGPT as an assistant in collaborative coding remains largely unexplored. In this paper, we analyze a dataset of 210 and 370 developers shared conversations with ChatGPT in GitHub pull requests (PRs) and issues. We manually examined the content of the conversations and characterized the dynamics of the sharing behavior, i.e., understanding the rationale behind the sharing, identifying the locations where the conversations were shared, and determining the roles of the developers who shared them. Our main observations are: (1) Developers seek ChatGPT assistance across 16 types of software engineering inquiries. In both conversations shared in PRs and issues, the most frequently encountered inquiry categories include code generation, conceptual questions, how-to guides, issue resolution, and code review. (2) Developers frequently engage with ChatGPT via multi-turn conversations where each prompt can fulfill various roles, such as unveiling initial or new tasks, iterative follow-up, and prompt refinement. Multi-turn conversations account for 33.2% of the conversations shared in PRs and 36.9% in issues. (3) In collaborative coding, developers leverage shared conversations with ChatGPT to facilitate their role-specific contributions, whether as authors of PRs or issues, code reviewers, or collaborators on issues. Our work serves as the first step towards understanding the dynamics between developers and ChatGPT in collaborative software development and opens up new directions for future research on the topic.
Sloan Digital Sky Survey IV: Mapping the Milky Way, Nearby Galaxies, and the Distant Universe
We describe the Sloan Digital Sky Survey IV (SDSS-IV), a project encompassing three major spectroscopic programs. The Apache Point Observatory Galactic Evolution Experiment 2 (APOGEE-2) is observing hundreds of thousands of Milky Way stars at high resolution and high signal-to-noise ratio in the near-infrared. The Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey is obtaining spatially-resolved spectroscopy for thousands of nearby galaxies (median redshift of z = 0.03). The extended Baryon Oscillation Spectroscopic Survey (eBOSS) is mapping the galaxy, quasar, and neutral gas distributions between redshifts z = 0.6 and 3.5 to constrain cosmology using baryon acoustic oscillations, redshift space distortions, and the shape of the power spectrum. Within eBOSS, we are conducting two major subprograms: the SPectroscopic IDentification of eROSITA Sources (SPIDERS), investigating X-ray AGN and galaxies in X-ray clusters, and the Time Domain Spectroscopic Survey (TDSS), obtaining spectra of variable sources. All programs use the 2.5-meter Sloan Foundation Telescope at Apache Point Observatory; observations there began in Summer 2014. APOGEE-2 also operates a second near-infrared spectrograph at the 2.5-meter du Pont Telescope at Las Campanas Observatory, with observations beginning in early 2017. Observations at both facilities are scheduled to continue through 2020. In keeping with previous SDSS policy, SDSS-IV provides regularly scheduled public data releases; the first one, Data Release 13, was made available in July 2016.
Barack's Wife Hillary: Using Knowledge-Graphs for Fact-Aware Language Modeling
Modeling human language requires the ability to not only generate fluent text but also encode factual knowledge. However, traditional language models are only capable of remembering facts seen at training time, and often have difficulty recalling them. To address this, we introduce the knowledge graph language model (KGLM), a neural language model with mechanisms for selecting and copying facts from a knowledge graph that are relevant to the context. These mechanisms enable the model to render information it has never seen before, as well as generate out-of-vocabulary tokens. We also introduce the Linked WikiText-2 dataset, a corpus of annotated text aligned to the Wikidata knowledge graph whose contents (roughly) match the popular WikiText-2 benchmark. In experiments, we demonstrate that the KGLM achieves significantly better performance than a strong baseline language model. We additionally compare different language model's ability to complete sentences requiring factual knowledge, showing that the KGLM outperforms even very large language models in generating facts.
Generating and Evaluating Tests for K-12 Students with Language Model Simulations: A Case Study on Sentence Reading Efficiency
Developing an educational test can be expensive and time-consuming, as each item must be written by experts and then evaluated by collecting hundreds of student responses. Moreover, many tests require multiple distinct sets of questions administered throughout the school year to closely monitor students' progress, known as parallel tests. In this study, we focus on tests of silent sentence reading efficiency, used to assess students' reading ability over time. To generate high-quality parallel tests, we propose to fine-tune large language models (LLMs) to simulate how previous students would have responded to unseen items. With these simulated responses, we can estimate each item's difficulty and ambiguity. We first use GPT-4 to generate new test items following a list of expert-developed rules and then apply a fine-tuned LLM to filter the items based on criteria from psychological measurements. We also propose an optimal-transport-inspired technique for generating parallel tests and show the generated tests closely correspond to the original test's difficulty and reliability based on crowdworker responses. Our evaluation of a generated test with 234 students from grades 2 to 8 produces test scores highly correlated (r=0.93) to those of a standard test form written by human experts and evaluated across thousands of K-12 students.
The Sharpness Disparity Principle in Transformers for Accelerating Language Model Pre-Training
Transformers consist of diverse building blocks, such as embedding layers, normalization layers, self-attention mechanisms, and point-wise feedforward networks. Thus, understanding the differences and interactions among these blocks is important. In this paper, we uncover a clear Sharpness Disparity across these blocks, which emerges early in training and intriguingly persists throughout the training process. Motivated by this finding, we propose Blockwise Learning Rate (LR), a strategy that tailors the LR to each block's sharpness, accelerating large language model (LLM) pre-training. By integrating Blockwise LR into AdamW, we consistently achieve lower terminal loss and nearly 2times speedup compared to vanilla AdamW. We demonstrate this acceleration across GPT-2 and LLaMA, with model sizes ranging from 0.12B to 1.1B and datasets of OpenWebText and MiniPile. Finally, we incorporate Blockwise LR into Adam-mini (Zhang et al., 2024), a recently proposed memory-efficient variant of Adam, achieving a combined 2times speedup and 2times memory saving. These results underscore the potential of exploiting the sharpness disparity to improve LLM training.
Efficient Memory Management for Large Language Model Serving with PagedAttention
High throughput serving of large language models (LLMs) requires batching sufficiently many requests at a time. However, existing systems struggle because the key-value cache (KV cache) memory for each request is huge and grows and shrinks dynamically. When managed inefficiently, this memory can be significantly wasted by fragmentation and redundant duplication, limiting the batch size. To address this problem, we propose PagedAttention, an attention algorithm inspired by the classical virtual memory and paging techniques in operating systems. On top of it, we build vLLM, an LLM serving system that achieves (1) near-zero waste in KV cache memory and (2) flexible sharing of KV cache within and across requests to further reduce memory usage. Our evaluations show that vLLM improves the throughput of popular LLMs by 2-4times with the same level of latency compared to the state-of-the-art systems, such as FasterTransformer and Orca. The improvement is more pronounced with longer sequences, larger models, and more complex decoding algorithms. vLLM's source code is publicly available at https://github.com/vllm-project/vllm
From Unimodal to Multimodal: Scaling up Projectors to Align Modalities
Recent contrastive multimodal vision-language models like CLIP have demonstrated robust open-world semantic understanding, becoming the standard image backbones for vision-language applications due to their aligned latent space. However, this practice has left powerful unimodal encoders for both vision and language underutilized in multimodal applications which raises a key question: Is there a plausible way to connect unimodal backbones for zero-shot vision-language tasks? To this end, we propose a novel approach that aligns vision and language modalities using only projection layers on pretrained, frozen unimodal encoders. Our method exploits the high semantic similarity between embedding spaces of well-trained vision and language models. It involves selecting semantically similar encoders in the latent space, curating a concept-rich dataset of image-caption pairs, and training simple MLP projectors. We evaluated our approach on 12 zero-shot classification datasets and 2 image-text retrieval datasets. Our best model, utilizing DINOv2 and All-Roberta-Large text encoder, achieves 76\(\%\) accuracy on ImageNet with a 20-fold reduction in data and 65 fold reduction in compute requirements. The proposed framework enhances the accessibility of model development while enabling flexible adaptation across diverse scenarios, offering an efficient approach to building multimodal models by utilizing existing unimodal architectures. Code and datasets will be released soon.
CSTS: Conditional Semantic Textual Similarity
Semantic textual similarity (STS) has been a cornerstone task in NLP that measures the degree of similarity between a pair of sentences, with applications in information retrieval, question answering, and embedding methods. However, it is an inherently ambiguous task, with the sentence similarity depending on the specific aspect of interest. We resolve this ambiguity by proposing a novel task called conditional STS (C-STS) which measures similarity conditioned on an aspect elucidated in natural language (hereon, condition). As an example, the similarity between the sentences "The NBA player shoots a three-pointer." and "A man throws a tennis ball into the air to serve." is higher for the condition "The motion of the ball." (both upward) and lower for "The size of the ball." (one large and one small). C-STS's advantages are two-fold: (1) it reduces the subjectivity and ambiguity of STS, and (2) enables fine-grained similarity evaluation using diverse conditions. C-STS contains almost 20,000 instances from diverse domains and we evaluate several state-of-the-art models to demonstrate that even the most performant fine-tuning and in-context learning models (GPT-4, Flan, SimCSE) find it challenging, with Spearman correlation scores of <50. We encourage the community to evaluate their models on C-STS to provide a more holistic view of semantic similarity and natural language understanding.
DataMUX: Data Multiplexing for Neural Networks
In this paper, we introduce data multiplexing (DataMUX), a technique that enables deep neural networks to process multiple inputs simultaneously using a single compact representation. DataMUX demonstrates that neural networks are capable of generating accurate predictions over mixtures of inputs, resulting in increased throughput with minimal extra memory requirements. Our approach uses two key components -- 1) a multiplexing layer that performs a fixed linear transformation to each input before combining them to create a mixed representation of the same size as a single input, which is then processed by the base network, and 2) a demultiplexing layer that converts the base network's output back into independent representations before producing predictions for each input. We show the viability of DataMUX for different architectures (Transformers, and to a lesser extent MLPs and CNNs) across six different tasks spanning sentence classification, named entity recognition and image classification. For instance, DataMUX for Transformers can multiplex up to 20x/40x inputs, achieving 11x/18x increase in throughput with minimal absolute performance drops of <2% and <4% respectively on MNLI, a natural language inference task. We also provide a theoretical construction for multiplexing in self-attention networks and analyze the effect of various design elements in DataMUX.
GatorTron: A Large Clinical Language Model to Unlock Patient Information from Unstructured Electronic Health Records
There is an increasing interest in developing artificial intelligence (AI) systems to process and interpret electronic health records (EHRs). Natural language processing (NLP) powered by pretrained language models is the key technology for medical AI systems utilizing clinical narratives. However, there are few clinical language models, the largest of which trained in the clinical domain is comparatively small at 110 million parameters (compared with billions of parameters in the general domain). It is not clear how large clinical language models with billions of parameters can help medical AI systems utilize unstructured EHRs. In this study, we develop from scratch a large clinical language model - GatorTron - using >90 billion words of text (including >82 billion words of de-identified clinical text) and systematically evaluate it on 5 clinical NLP tasks including clinical concept extraction, medical relation extraction, semantic textual similarity, natural language inference (NLI), and medical question answering (MQA). We examine how (1) scaling up the number of parameters and (2) scaling up the size of the training data could benefit these NLP tasks. GatorTron models scale up the clinical language model from 110 million to 8.9 billion parameters and improve 5 clinical NLP tasks (e.g., 9.6% and 9.5% improvement in accuracy for NLI and MQA), which can be applied to medical AI systems to improve healthcare delivery. The GatorTron models are publicly available at: https://catalog.ngc.nvidia.com/orgs/nvidia/teams/clara/models/gatortron_og.
The Apache Point Observatory Galactic Evolution Experiment (APOGEE) Spectrographs
We describe the design and performance of the near-infrared (1.51--1.70 micron), fiber-fed, multi-object (300 fibers), high resolution (R = lambda/delta lambda ~ 22,500) spectrograph built for the Apache Point Observatory Galactic Evolution Experiment (APOGEE). APOGEE is a survey of ~ 10^5 red giant stars that systematically sampled all Milky Way populations (bulge, disk, and halo) to study the Galaxy's chemical and kinematical history. It was part of the Sloan Digital Sky Survey III (SDSS-III) from 2011 -- 2014 using the 2.5 m Sloan Foundation Telescope at Apache Point Observatory, New Mexico. The APOGEE-2 survey is now using the spectrograph as part of SDSS-IV, as well as a second spectrograph, a close copy of the first, operating at the 2.5 m du Pont Telescope at Las Campanas Observatory in Chile. Although several fiber-fed, multi-object, high resolution spectrographs have been built for visual wavelength spectroscopy, the APOGEE spectrograph is one of the first such instruments built for observations in the near-infrared. The instrument's successful development was enabled by several key innovations, including a "gang connector" to allow simultaneous connections of 300 fibers; hermetically sealed feedthroughs to allow fibers to pass through the cryostat wall continuously; the first cryogenically deployed mosaic volume phase holographic grating; and a large refractive camera that includes mono-crystalline silicon and fused silica elements with diameters as large as ~ 400 mm. This paper contains a comprehensive description of all aspects of the instrument including the fiber system, optics and opto-mechanics, detector arrays, mechanics and cryogenics, instrument control, calibration system, optical performance and stability, lessons learned, and design changes for the second instrument.
RouteLLM: Learning to Route LLMs with Preference Data
Large language models (LLMs) exhibit impressive capabilities across a wide range of tasks, yet the choice of which model to use often involves a trade-off between performance and cost. More powerful models, though effective, come with higher expenses, while less capable models are more cost-effective. To address this dilemma, we propose several efficient router models that dynamically select between a stronger and a weaker LLM during inference, aiming to optimize the balance between cost and response quality. We develop a training framework for these routers leveraging human preference data and data augmentation techniques to enhance performance. Our evaluation on widely-recognized benchmarks shows that our approach significantly reduces costs-by over 2 times in certain cases-without compromising the quality of responses. Interestingly, our router models also demonstrate significant transfer learning capabilities, maintaining their performance even when the strong and weak models are changed at test time. This highlights the potential of these routers to provide a cost-effective yet high-performance solution for deploying LLMs.
Reconstructing unseen modalities and pathology with an efficient Recurrent Inference Machine
Objective: To allow efficient learning using the Recurrent Inference Machine (RIM) for image reconstruction whereas not being strictly dependent on the training data distribution so that unseen modalities and pathologies are still accurately recovered. Methods: Theoretically, the RIM learns to solve the inverse problem of accelerated-MRI reconstruction whereas being robust to variable imaging conditions. The efficiency and generalization capabilities with different training datasets were studied, as well as recurrent network units with decreasing complexity: the Gated Recurrent Unit (GRU), the Minimal Gated Unit (MGU), and the Independently Recurrent Neural Network (IndRNN), to reduce inference times. Validation was performed against Compressed Sensing (CS) and further assessed based on data unseen during training. A pathology study was conducted by reconstructing simulated white matter lesions and prospectively undersampled data of a Multiple Sclerosis patient. Results: Training on a single modality of 3T T_1-weighted brain data appeared sufficient to also reconstruct 7T T_{2}^*-weighted brain and 3T T_2-weighted knee data. The IndRNN is an efficient recurrent unit, reducing inference time by 68\% compared to CS, whereas maintaining performance. The RIM was able to reconstruct lesions unseen during training more accurately than CS when trained on T_2-weighted knee data. Training on T_1-weighted brain data and on combined data slightly enhanced the signal compared to CS. Conclusion: The RIM is efficient when decreasing its complexity, which reduces the inference time, whereas still being able to reconstruct data and pathology that was unseen during training.
On the Importance of Text Preprocessing for Multimodal Representation Learning and Pathology Report Generation
Vision-language models in pathology enable multimodal case retrieval and automated report generation. Many of the models developed so far, however, have been trained on pathology reports that include information which cannot be inferred from paired whole slide images (e.g., patient history), potentially leading to hallucinated sentences in generated reports. To this end, we investigate how the selection of information from pathology reports for vision-language modeling affects the quality of the multimodal representations and generated reports. More concretely, we compare a model trained on full reports against a model trained on preprocessed reports that only include sentences describing the cell and tissue appearances based on the H&E-stained slides. For the experiments, we built upon the BLIP-2 framework and used a cutaneous melanocytic lesion dataset of 42,433 H&E-stained whole slide images and 19,636 corresponding pathology reports. Model performance was assessed using image-to-text and text-to-image retrieval, as well as qualitative evaluation of the generated reports by an expert pathologist. Our results demonstrate that text preprocessing prevents hallucination in report generation. Despite the improvement in the quality of the generated reports, training the vision-language model on full reports showed better cross-modal retrieval performance.
AI for operational methane emitter monitoring from space
Mitigating methane emissions is the fastest way to stop global warming in the short-term and buy humanity time to decarbonise. Despite the demonstrated ability of remote sensing instruments to detect methane plumes, no system has been available to routinely monitor and act on these events. We present MARS-S2L, an automated AI-driven methane emitter monitoring system for Sentinel-2 and Landsat satellite imagery deployed operationally at the United Nations Environment Programme's International Methane Emissions Observatory. We compile a global dataset of thousands of super-emission events for training and evaluation, demonstrating that MARS-S2L can skillfully monitor emissions in a diverse range of regions globally, providing a 216% improvement in mean average precision over a current state-of-the-art detection method. Running this system operationally for six months has yielded 457 near-real-time detections in 22 different countries of which 62 have already been used to provide formal notifications to governments and stakeholders.
Post-Training Sparse Attention with Double Sparsity
The inference process for large language models is slow and memory-intensive, with one of the most critical bottlenecks being excessive Key-Value (KV) cache accesses. This paper introduces "Double Sparsity," a novel post-training sparse attention technique designed to alleviate this bottleneck by reducing KV cache access. Double Sparsity combines token sparsity, which focuses on utilizing only the important tokens for computing self-attention, with channel sparsity, an approach that uses important feature channels for identifying important tokens. Our key insight is that the pattern of channel sparsity is relatively static, allowing us to use offline calibration to make it efficient at runtime, thereby enabling accurate and efficient identification of important tokens. Moreover, this method can be combined with offloading to achieve significant memory usage reduction. Experimental results demonstrate that Double Sparsity can achieve 1{16} token and channel sparsity with minimal impact on accuracy across various tasks, including wiki-2 perplexity, key-value retrieval, and long context benchmarks with models including Llama-2-7B, Llama-2-70B, and Mixtral-8x7B. It brings up to a 14.1times acceleration in attention operations and a 1.9times improvement in end-to-end inference on GPUs. With offloading, it achieves a decoding speed acceleration of 16.3times compared to state-of-the-art solutions at a sequence length of 256K. Our code is publicly available at https://github.com/andy-yang-1/DoubleSparse.
ACES: Automatic Cohort Extraction System for Event-Stream Datasets
Reproducibility remains a significant challenge in machine learning (ML) for healthcare. In this field, datasets, model pipelines, and even task/cohort definitions are often private, leading to a significant barrier in sharing, iterating, and understanding ML results on electronic health record (EHR) datasets. In this paper, we address a significant part of this problem by introducing the Automatic Cohort Extraction System for Event-Stream Datasets (ACES). This tool is designed to simultaneously simplify the development of task/cohorts for ML in healthcare and enable the reproduction of these cohorts, both at an exact level for single datasets and at a conceptual level across datasets. To accomplish this, ACES provides (1) a highly intuitive and expressive configuration language for defining both dataset-specific concepts and dataset-agnostic inclusion/exclusion criteria, and (2) a pipeline to automatically extract patient records that meet these defined criteria from real-world data. ACES can be automatically applied to any dataset in either the Medical Event Data Standard (MEDS) or EventStreamGPT (ESGPT) formats, or to *any* dataset for which the necessary task-specific predicates can be extracted in an event-stream form. ACES has the potential to significantly lower the barrier to entry for defining ML tasks, redefine the way researchers interact with EHR datasets, and significantly improve the state of reproducibility for ML studies in this modality. ACES is available at https://github.com/justin13601/aces.
Constraints on the variation of the fine-structure constant at 3<z<10 with JWST emission-line galaxies
We present constraints on the spacetime variation of the fine-structure constant alpha at redshifts 2.5le z<9.5 using JWST emission-line galaxies. The galaxy sample consists of 621 high-quality spectra with strong and narrow [O III] lambdalambda4959,5007 doublet emission lines from 578 galaxies, including 232 spectra at z>5. The [O III] doublet lines are arguably the best emission lines to probe the variation in alpha. We divide our sample into six subsamples based on redshift and calculate the relative variation Deltaalpha/alpha for the individual subsamples. The calculated Deltaalpha/alpha values are consistent with zero within 1sigma at all redshifts, suggesting no time variation in alpha above a level of (1-2) times10^{-4} (1sigma) in the past 13.2 billion years. When the whole sample is combined, the constraint is improved to be Deltaalpha/alpha = (0.2pm0.7) times10^{-4}. We further test the spatial variation in alpha using four subsamples of galaxies in four different directions on the sky. The measured Deltaalpha/alpha values are consistent with zero at a 1sigma level of sim 2times10^{-4}. While the constraints in this work are not as stringent as those from lower-redshift quasar absorption lines in previous studies, this work uses an independent tracer and provides the first constraints on Deltaalpha/alpha at the highest redshifts. With the growing number of emission-line galaxies from JWST, we expect to achieve stronger constraints in the future.
Incremental Semi-supervised Federated Learning for Health Inference via Mobile Sensing
Mobile sensing appears as a promising solution for health inference problem (e.g., influenza-like symptom recognition) by leveraging diverse smart sensors to capture fine-grained information about human behaviors and ambient contexts. Centralized training of machine learning models can place mobile users' sensitive information under privacy risks due to data breach and misexploitation. Federated Learning (FL) enables mobile devices to collaboratively learn global models without the exposure of local private data. However, there are challenges of on-device FL deployment using mobile sensing: 1) long-term and continuously collected mobile sensing data may exhibit domain shifts as sensing objects (e.g. humans) have varying behaviors as a result of internal and/or external stimulus; 2) model retraining using all available data may increase computation and memory burden; and 3) the sparsity of annotated crowd-sourced data causes supervised FL to lack robustness. In this work, we propose FedMobile, an incremental semi-supervised federated learning algorithm, to train models semi-supervisedly and incrementally in a decentralized online fashion. We evaluate FedMobile using a real-world mobile sensing dataset for influenza-like symptom recognition. Our empirical results show that FedMobile-trained models achieve the best results in comparison to the selected baseline methods.
System Design for an Integrated Lifelong Reinforcement Learning Agent for Real-Time Strategy Games
As Artificial and Robotic Systems are increasingly deployed and relied upon for real-world applications, it is important that they exhibit the ability to continually learn and adapt in dynamically-changing environments, becoming Lifelong Learning Machines. Continual/lifelong learning (LL) involves minimizing catastrophic forgetting of old tasks while maximizing a model's capability to learn new tasks. This paper addresses the challenging lifelong reinforcement learning (L2RL) setting. Pushing the state-of-the-art forward in L2RL and making L2RL useful for practical applications requires more than developing individual L2RL algorithms; it requires making progress at the systems-level, especially research into the non-trivial problem of how to integrate multiple L2RL algorithms into a common framework. In this paper, we introduce the Lifelong Reinforcement Learning Components Framework (L2RLCF), which standardizes L2RL systems and assimilates different continual learning components (each addressing different aspects of the lifelong learning problem) into a unified system. As an instantiation of L2RLCF, we develop a standard API allowing easy integration of novel lifelong learning components. We describe a case study that demonstrates how multiple independently-developed LL components can be integrated into a single realized system. We also introduce an evaluation environment in order to measure the effect of combining various system components. Our evaluation environment employs different LL scenarios (sequences of tasks) consisting of Starcraft-2 minigames and allows for the fair, comprehensive, and quantitative comparison of different combinations of components within a challenging common evaluation environment.
ReCapture: Generative Video Camera Controls for User-Provided Videos using Masked Video Fine-Tuning
Recently, breakthroughs in video modeling have allowed for controllable camera trajectories in generated videos. However, these methods cannot be directly applied to user-provided videos that are not generated by a video model. In this paper, we present ReCapture, a method for generating new videos with novel camera trajectories from a single user-provided video. Our method allows us to re-generate the reference video, with all its existing scene motion, from vastly different angles and with cinematic camera motion. Notably, using our method we can also plausibly hallucinate parts of the scene that were not observable in the reference video. Our method works by (1) generating a noisy anchor video with a new camera trajectory using multiview diffusion models or depth-based point cloud rendering and then (2) regenerating the anchor video into a clean and temporally consistent reangled video using our proposed masked video fine-tuning technique.
S*: Test Time Scaling for Code Generation
Increasing test-time compute for LLMs shows promise across domains but remains underexplored in code generation, despite extensive study in math. In this paper, we propose S*, the first hybrid test-time scaling framework that substantially improves the coverage and selection accuracy of generated code. S* extends the existing parallel scaling paradigm with sequential scaling to push performance boundaries. It further leverages a novel selection mechanism that adaptively generates distinguishing inputs for pairwise comparison, combined with execution-grounded information to robustly identify correct solutions. We evaluate across 12 Large Language Models and Large Reasoning Model and show: (1) S* consistently improves performance across model families and sizes, enabling a 3B model to outperform GPT-4o-mini; (2) S* enables non-reasoning models to surpass reasoning models - GPT-4o-mini with S* outperforms o1-preview by 3.7% on LiveCodeBench; (3) S* further boosts state-of-the-art reasoning models - DeepSeek-R1-Distill-Qwen-32B with S* achieves 85.7% on LiveCodeBench, approaching o1 (high) at 88.5%. Code will be available under https://github.com/NovaSky-AI/SkyThought.
Foundation Models for Generalist Geospatial Artificial Intelligence
Significant progress in the development of highly adaptable and reusable Artificial Intelligence (AI) models is expected to have a significant impact on Earth science and remote sensing. Foundation models are pre-trained on large unlabeled datasets through self-supervision, and then fine-tuned for various downstream tasks with small labeled datasets. This paper introduces a first-of-a-kind framework for the efficient pre-training and fine-tuning of foundational models on extensive geospatial data. We have utilized this framework to create Prithvi, a transformer-based geospatial foundational model pre-trained on more than 1TB of multispectral satellite imagery from the Harmonized Landsat-Sentinel 2 (HLS) dataset. Our study demonstrates the efficacy of our framework in successfully fine-tuning Prithvi to a range of Earth observation tasks that have not been tackled by previous work on foundation models involving multi-temporal cloud gap imputation, flood mapping, wildfire scar segmentation, and multi-temporal crop segmentation. Our experiments show that the pre-trained model accelerates the fine-tuning process compared to leveraging randomly initialized weights. In addition, pre-trained Prithvi compares well against the state-of-the-art, e.g., outperforming a conditional GAN model in multi-temporal cloud imputation by up to 5pp (or 5.7%) in the structural similarity index. Finally, due to the limited availability of labeled data in the field of Earth observation, we gradually reduce the quantity of available labeled data for refining the model to evaluate data efficiency and demonstrate that data can be decreased significantly without affecting the model's accuracy. The pre-trained 100 million parameter model and corresponding fine-tuning workflows have been released publicly as open source contributions to the global Earth sciences community through Hugging Face.
LightSeq: Sequence Level Parallelism for Distributed Training of Long Context Transformers
Increasing the context length of large language models (LLMs) unlocks fundamentally new capabilities, but also significantly increases the memory footprints of training. Previous model-parallel systems such as Megatron-LM partition and compute different attention heads in parallel, resulting in large communication volumes, so they cannot scale beyond the number of attention heads, thereby hindering its adoption. In this paper, we introduce a new approach, LightSeq, for long-context LLMs training. LightSeq has many notable advantages. First, LightSeq partitions over the sequence dimension, hence is agnostic to model architectures and readily applicable for models with varying numbers of attention heads, such as Multi-Head, Multi-Query and Grouped-Query attention. Second, LightSeq not only requires up to 4.7x less communication than Megatron-LM on popular LLMs but also overlaps the communication with computation. To further reduce the training time, LightSeq features a novel gradient checkpointing scheme to bypass an forward computation for memory-efficient attention. We evaluate LightSeq on Llama-7B and its variants with sequence lengths from 32K to 512K. Through comprehensive experiments on single and cross-node training, we show that LightSeq achieves up to 1.24-2.01x end-to-end speedup, and a 2-8x longer sequence length on models with fewer heads, compared to Megatron-LM. Codes will be available at https://github.com/RulinShao/LightSeq.
INQUIRE: A Natural World Text-to-Image Retrieval Benchmark
We introduce INQUIRE, a text-to-image retrieval benchmark designed to challenge multimodal vision-language models on expert-level queries. INQUIRE includes iNaturalist 2024 (iNat24), a new dataset of five million natural world images, along with 250 expert-level retrieval queries. These queries are paired with all relevant images comprehensively labeled within iNat24, comprising 33,000 total matches. Queries span categories such as species identification, context, behavior, and appearance, emphasizing tasks that require nuanced image understanding and domain expertise. Our benchmark evaluates two core retrieval tasks: (1) INQUIRE-Fullrank, a full dataset ranking task, and (2) INQUIRE-Rerank, a reranking task for refining top-100 retrievals. Detailed evaluation of a range of recent multimodal models demonstrates that INQUIRE poses a significant challenge, with the best models failing to achieve an mAP@50 above 50%. In addition, we show that reranking with more powerful multimodal models can enhance retrieval performance, yet there remains a significant margin for improvement. By focusing on scientifically-motivated ecological challenges, INQUIRE aims to bridge the gap between AI capabilities and the needs of real-world scientific inquiry, encouraging the development of retrieval systems that can assist with accelerating ecological and biodiversity research. Our dataset and code are available at https://inquire-benchmark.github.io
ALMA Lensing Cluster Survey: Physical characterization of near-infrared-dark intrinsically faint ALMA sources at z=2-4
We present results from Atacama Large Millimeter/submillimeter Array (ALMA) spectral line-scan observations at 3-mm and 2-mm bands of three near-infrared-dark (NIR-dark) galaxies behind two massive lensing clusters MACS J0417.5-1154 and RXC J0032.1+1808. Each of these three sources is a faint (de-lensed S_{1.2 mm} < 1 mJy) triply lensed system originally discovered in the ALMA Lensing Cluster Survey. We have successfully detected CO and [C I] emission lines and confirmed that their spectroscopic redshifts are z=3.652, 2.391, and 2.985. By utilizing a rich multi-wavelength data set, we find that the NIR-dark galaxies are located on the star formation main sequence in the intrinsic stellar mass range of log (M_*/M_odot) = 9.8 - 10.4, which is about one order of magnitude lower than that of typical submillimeter galaxies (SMGs). These NIR-dark galaxies show a variety in gas depletion times and spatial extent of dust emission. One of the three is a normal star-forming galaxy with gas depletion time consistent with a scaling relation, and its infrared surface brightness is an order of magnitude smaller than that of typical SMGs. Since this galaxy has an elongated axis ratio of sim 0.17, we argue that normal star-forming galaxies in an edge-on configuration can be heavily dust-obscured. This implies that existing deep WFC3/F160W surveys may miss a fraction of typical star-forming main-sequence galaxies due to their edge-on orientation.
Biomedical knowledge graph-optimized prompt generation for large language models
Large Language Models (LLMs) are being adopted at an unprecedented rate, yet still face challenges in knowledge-intensive domains like biomedicine. Solutions such as pre-training and domain-specific fine-tuning add substantial computational overhead, requiring further domain expertise. Here, we introduce a token-optimized and robust Knowledge Graph-based Retrieval Augmented Generation (KG-RAG) framework by leveraging a massive biomedical KG (SPOKE) with LLMs such as Llama-2-13b, GPT-3.5-Turbo and GPT-4, to generate meaningful biomedical text rooted in established knowledge. Compared to the existing RAG technique for Knowledge Graphs, the proposed method utilizes minimal graph schema for context extraction and uses embedding methods for context pruning. This optimization in context extraction results in more than 50% reduction in token consumption without compromising the accuracy, making a cost-effective and robust RAG implementation on proprietary LLMs. KG-RAG consistently enhanced the performance of LLMs across diverse biomedical prompts by generating responses rooted in established knowledge, accompanied by accurate provenance and statistical evidence (if available) to substantiate the claims. Further benchmarking on human curated datasets, such as biomedical true/false and multiple-choice questions (MCQ), showed a remarkable 71% boost in the performance of the Llama-2 model on the challenging MCQ dataset, demonstrating the framework's capacity to empower open-source models with fewer parameters for domain specific questions. Furthermore, KG-RAG enhanced the performance of proprietary GPT models, such as GPT-3.5 and GPT-4. In summary, the proposed framework combines explicit and implicit knowledge of KG and LLM in a token optimized fashion, thus enhancing the adaptability of general-purpose LLMs to tackle domain-specific questions in a cost-effective fashion.
Message in a Bottle -- An Update to the Golden Record
In this first part of our series, we delve into the foundational aspects of the "Message in a Bottle" (henceforth referred to as MIAB). This study stands as a continuation of the legacy set by the Voyager Golden Records launched aboard Voyager 1 and 2 in 1977, which aimed to communicate with intelligent species beyond our world. These Records continue to serve not only as a snapshot of Earth and humanity but also carry forth our desire for establishing contact with advanced alien civilizations. Given the absence of mutually understood signs, symbols, and semiotic conventions, MIAB, like its predecessor, seeks to use scientific methods to design an innovative means of communication encapsulating the story of humanity. Our aim is to convey our collective knowledge, feelings, innovations, and aspirations in a manner that offers a universal, yet contextual understanding of human society, the evolution of life on Earth, and our hopes and concerns for the future. Through this time and space traveling capsule, we also strive to inspire and unify current and future generations to celebrate and safeguard our shared human experience.
Deep Reinforcement Learning for Conservation Decisions
Can machine learning help us make better decisions about a changing planet? In this paper, we illustrate and discuss the potential of a promising corner of machine learning known as _reinforcement learning_ (RL) to help tackle the most challenging conservation decision problems. RL is uniquely well suited to conservation and global change challenges for three reasons: (1) RL explicitly focuses on designing an agent who _interacts_ with an environment which is dynamic and uncertain, (2) RL approaches do not require massive amounts of data, (3) RL approaches would utilize rather than replace existing models, simulations, and the knowledge they contain. We provide a conceptual and technical introduction to RL and its relevance to ecological and conservation challenges, including examples of a problem in setting fisheries quotas and in managing ecological tipping points. Four appendices with annotated code provide a tangible introduction to researchers looking to adopt, evaluate, or extend these approaches.
Prithvi WxC: Foundation Model for Weather and Climate
Triggered by the realization that AI emulators can rival the performance of traditional numerical weather prediction models running on HPC systems, there is now an increasing number of large AI models that address use cases such as forecasting, downscaling, or nowcasting. While the parallel developments in the AI literature focus on foundation models -- models that can be effectively tuned to address multiple, different use cases -- the developments on the weather and climate side largely focus on single-use cases with particular emphasis on mid-range forecasting. We close this gap by introducing Prithvi WxC, a 2.3 billion parameter foundation model developed using 160 variables from the Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2). Prithvi WxC employs an encoder-decoder-based architecture, incorporating concepts from various recent transformer models to effectively capture both regional and global dependencies in the input data. The model has been designed to accommodate large token counts to model weather phenomena in different topologies at fine resolutions. Furthermore, it is trained with a mixed objective that combines the paradigms of masked reconstruction with forecasting. We test the model on a set of challenging downstream tasks namely: Autoregressive rollout forecasting, Downscaling, Gravity wave flux parameterization, and Extreme events estimation. The pretrained model with 2.3 billion parameters, along with the associated fine-tuning workflows, has been publicly released as an open-source contribution via Hugging Face.
LEANN: A Low-Storage Vector Index
Embedding-based search is widely used in applications such as recommendation and retrieval-augmented generation (RAG). Recently, there is a growing demand to support these capabilities over personal data stored locally on devices. However, maintaining the necessary data structure associated with the embedding-based search is often infeasible due to its high storage overhead. For example, indexing 100 GB of raw data requires 150 to 700 GB of storage, making local deployment impractical. Reducing this overhead while maintaining search quality and latency becomes a critical challenge. In this paper, we present LEANN, a storage-efficient approximate nearest neighbor (ANN) search index optimized for resource-constrained personal devices. LEANN combines a compact graph-based structure with an efficient on-the-fly recomputation strategy to enable fast and accurate retrieval with minimal storage overhead. Our evaluation shows that LEANN reduces index size to under 5% of the original raw data, achieving up to 50 times smaller storage than standard indexes, while maintaining 90% top-3 recall in under 2 seconds on real-world question answering benchmarks.
Surprising Variation of Gamma Rays from the Sun over the Solar Cycle Revealed with Fermi-LAT
The steady-state gamma-ray emission from the Sun is thought to consist of two emission components due to interactions with Galactic cosmic rays: (1) a hadronic component covering the solar disk, and (2) a leptonic component peaking at the solar edge and extending into the heliosphere. The flux of these components is expected to vary with the 11-year solar cycle, being highest during solar minimum and lowest during solar maximum, because it is correlated with the cosmic-ray flux. No study has yet analyzed the flux variation of the two components separately over solar cycles. In this work, we measure the temporal variations of the flux of each component over 15 years of Fermi Large Area Telescope observations and compare them with the sunspot number and Galactic cosmic-ray flux from AMS-02 near the Earth. We find that the flux variation of the disk anticorrelates with solar activity and correlates with cosmic-ray protons, confirming its emission mechanism. The flux variation of the extended component anticorrelates with solar activity only until mid 2012. After that, we no longer observe any correlation or anticorrelation, even with the CR electron flux. This most likely suggests that cosmic-ray transport and modulation in the inner heliosphere are unexpectedly complex and different for electrons and protons or, alternatively, the presence of an additional, unknown component of gamma rays or cosmic rays. These findings impact space weather research and emphasize the need for close monitoring of Cycle 25 and the ongoing polarity reversal.
UNIONS: The Ultraviolet Near-Infrared Optical Northern Survey
The Ultraviolet Near-Infrared Optical Northern Survey (UNIONS) is a "collaboration of collaborations" that is using the Canada-France-Hawai'i Telescope, the Pan-STARRS telescopes, and the Subaru Observatory to obtain ugriz images of a core survey region of 6250 deg^2 of the northern sky. The 10sigma point source depth of the data, as measured within a 2-arcsecond diameter aperture, are [u,g,r,i,z] = [23.7, 24.5, 24.2, 23.8, 23.3]\ in AB magnitudes. UNIONS is addressing some of the most fundamental questions in astronomy, including the properties of dark matter, the growth of structure in the Universe from the very smallest galaxies to large-scale structure, and the assembly of the Milky Way. It is set to become the major ground-based legacy survey for the northern hemisphere for the next decade and provides an essential northern complement to the static-sky science of the Vera C. Rubin Observatory's Legacy Survey of Space and Time. UNIONS supports the core science mission of the {\it Euclid} space mission by providing the data necessary in the northern hemisphere for the calibration of the wavelength dependence of the {\it Euclid} point-spread function and derivation of photometric redshifts in the North Galactic Cap. This region contains the highest quality sky for {\it Euclid}, with low backgrounds from the zodiacal light, stellar density, extinction, and emission from Galactic cirrus. Here, we describe the UNIONS survey components, science goals, data products, and the current status of the overall program.
PDRs4All. XII. FUV-driven formation of hydrocarbon radicals and their relation with PAHs
We present subarcsecond-resolution ALMA mosaics of the Orion Bar PDR in [CI] 609 um, C2H (4-3), and C18O (3-2) emission lines, complemented by JWST images of H2 and aromatic infrared band (AIB) emission. The rim of the Bar shows very corrugated structures made of small-scale H2 dissociation fronts (DFs). The [CI] 609 um emission peaks very close (~0.002 pc) to the main H2-emitting DFs, suggesting the presence of gas density gradients. These DFs are also bright and remarkably similar in C2H emission, which traces 'hydrocarbon radical peaks' characterized by very high C2H abundances, reaching up to several x10^-7. The high abundance of C2H and of related hydrocarbon radicals, such as CH3, CH2, and CH, can be attributed to gas-phase reactions driven by elevated temperatures, the presence of C+ and C, and the reactivity of FUV-pumped H2. The hydrocarbon radical peaks roughly coincide with maxima of the 3.4/3.3 um AIB intensity ratio, a proxy for the aliphatic-to-aromatic content of PAHs. This implies that the conditions triggering the formation of simple hydrocarbons also favor the formation (and survival) of PAHs with aliphatic side groups, potentially via the contribution of bottom-up processes in which abundant hydrocarbon radicals react in situ with PAHs. Ahead of the DFs, in the atomic PDR zone (where [H]>>[H2]), the AIB emission is brightest, but small PAHs and carbonaceous grains undergo photo-processing due to the stronger FUV field. Our detection of trace amounts of C2H in this zone may result from the photoerosion of these species. This study provides a spatially resolved view of the chemical stratification of key carbon carriers in a PDR. Overall, both bottom-up and top-down processes appear to link simple hydrocarbon molecules with PAHs in molecular clouds; however, the exact chemical pathways and their relative contributions remain to be quantified.
A search for extremely-high-energy neutrinos and first constraints on the ultra-high-energy cosmic-ray proton fraction with IceCube
We present a search for the diffuse extremely-high-energy neutrino flux using 12.6 years of IceCube data. The non-observation of neutrinos with energies well above 10 , PeV constrains the all-flavor neutrino flux at 10^{18} , eV to a level of E^2 Phi_{nu_e + nu_mu + nu_tau} simeq 10^{-8} , GeV , cm^{-2} , s^{-1} , sr^{-1}, the most stringent limit to date. Using this data, we constrain the proton fraction of ultra-high-energy cosmic rays (UHECRs) above simeq 30 , EeV to be lesssim 70,% (at 90,% CL) if the cosmological evolution of the sources is comparable to or stronger than the star formation rate. This result complements direct air-shower measurements by being insensitive to uncertainties associated with hadronic interaction models. It is the first such result to disfavor the ``proton-only" hypothesis for UHECRs using neutrino data.
A region-wide, multi-year set of crop field boundary labels for Africa
African agriculture is undergoing rapid transformation. Annual maps of crop fields are key to understanding the nature of this transformation, but such maps are currently lacking and must be developed using advanced machine learning models trained on high resolution remote sensing imagery. To enable the development of such models, we delineated field boundaries in 33,746 Planet images captured between 2017 and 2023 across the continent using a custom labeling platform with built-in procedures for assessing and mitigating label error. We collected 42,403 labels, including 7,204 labels arising from tasks dedicated to assessing label quality (Class 1 labels), 32,167 from sites mapped once by a single labeller (Class 2) and 3,032 labels from sites where 3 or more labellers were tasked to map the same location (Class 4). Class 1 labels were used to calculate labeller-specific quality scores, while Class 1 and 4 sites mapped by at least 3 labellers were used to further evaluate label uncertainty using a Bayesian risk metric. Quality metrics showed that label quality was moderately high (0.75) for measures of total field extent, but low regarding the number of individual fields delineated (0.33), and the position of field edges (0.05). These values are expected when delineating small-scale fields in 3-5 m resolution imagery, which can be too coarse to reliably distinguish smaller fields, particularly in dense croplands, and therefore requires substantial labeller judgement. Nevertheless, previous work shows that such labels can train effective field mapping models. Furthermore, this large, probabilistic sample on its own provides valuable insight into regional agricultural characteristics, highlighting variations in the median field size and density. The imagery and vectorized labels along with quality information is available for download from two public repositories.
TorchMD-Net 2.0: Fast Neural Network Potentials for Molecular Simulations
Achieving a balance between computational speed, prediction accuracy, and universal applicability in molecular simulations has been a persistent challenge. This paper presents substantial advancements in the TorchMD-Net software, a pivotal step forward in the shift from conventional force fields to neural network-based potentials. The evolution of TorchMD-Net into a more comprehensive and versatile framework is highlighted, incorporating cutting-edge architectures such as TensorNet. This transformation is achieved through a modular design approach, encouraging customized applications within the scientific community. The most notable enhancement is a significant improvement in computational efficiency, achieving a very remarkable acceleration in the computation of energy and forces for TensorNet models, with performance gains ranging from 2-fold to 10-fold over previous iterations. Other enhancements include highly optimized neighbor search algorithms that support periodic boundary conditions and the smooth integration with existing molecular dynamics frameworks. Additionally, the updated version introduces the capability to integrate physical priors, further enriching its application spectrum and utility in research. The software is available at https://github.com/torchmd/torchmd-net.
An Atlas of Color-selected Quiescent Galaxies at $z>3$ in Public $JWST$ Fields
We present the results of a systematic search for candidate quiescent galaxies in the distant Universe in eleven JWST fields with publicly available observations collected during the first three months of operations and covering an effective sky area of sim145 arcmin^2. We homogeneously reduce the new JWST data and combine them with existing observations from the Hubble,Space,Telescope. We select a robust sample of sim80 candidate quiescent and quenching galaxies at 3 < z < 5 using two methods: (1) based on their rest-frame UVJ colors, and (2) a novel quantitative approach based on Gaussian Mixture Modeling of the NUV-U, U-V, and V-J rest-frame color space, which is more sensitive to recently quenched objects. We measure comoving number densities of massive (M_stargeq 10^{10.6} M_odot) quiescent galaxies consistent with previous estimates relying on ground-based observations, after homogenizing the results in the literature with our mass and redshift intervals. However, we find significant field-to-field variations of the number densities up to a factor of 2-3, highlighting the effect of cosmic variance and suggesting the presence of overdensities of red quiescent galaxies at z>3, as it could be expected for highly clustered massive systems. Importantly, JWST enables the robust identification of quenching/quiescent galaxy candidates at lower masses and higher redshifts than before, challenging standard formation scenarios. All data products, including the literature compilation, are made publicly available.
A 2.4% Determination of the Local Value of the Hubble Constant
We use the Wide Field Camera 3 (WFC3) on the Hubble Space Telescope (HST) to reduce the uncertainty in the local value of the Hubble constant (H_0) from 3.3% to 2.4%. Improvements come from new, near-infrared observations of Cepheid variables in 11 new hosts of recent SNe~Ia, more than doubling the sample of SNe~Ia having a Cepheid-calibrated distance for a total of 19; these leverage the magnitude-z relation based on 300 SNe~Ia at z<0.15. All 19 hosts and the megamaser system NGC4258 were observed with WFC3, thus nullifying cross-instrument zeropoint errors. Other improvements include a 33% reduction in the systematic uncertainty in the maser distance to NGC4258, more Cepheids and a more robust distance to the LMC from late-type DEBs, HST observations of Cepheids in M31, and new HST-based trigonometric parallaxes for Milky Way (MW) Cepheids. We consider four geometric distance calibrations of Cepheids: (i) megamasers in NGC4258, (ii) 8 DEBs in the LMC, (iii) 15 MW Cepheids with parallaxes, and (iv) 2 DEBs in M31. H_0 from each is 72.25+/-2.51, 72.04+/-2.67, 76.18+/-2.37, and 74.50+/-3.27 km/sec/Mpc, respectively. Our best estimate of 73.24+/-1.74 km/sec/Mpc combines the anchors NGC4258, MW, and LMC, and includes systematic errors for a final uncertainty of 2.4%. This value is 3.4 sigma higher than 66.93+/-0.62 km/sec/Mpc predicted by LambdaCDM with 3 neutrinos with mass 0.06 eV and the Planck data, but reduces to 2.1 sigma relative to the prediction of 69.3+/-0.7 km/sec/Mpc with the combination of WMAP+ACT+SPT+BAO, suggesting systematic uncertainties in CMB measurements may play a role in the tension. If we take the conflict between Planck and H_0 at face value, one plausible explanation could involve an additional source of dark radiation in the early Universe in the range of Delta N_eff=0.4-1. We anticipate significant improvements in H_0 from upcoming parallax measurements.
The Chandra Source Catalog
The Chandra Source Catalog (CSC) is a general purpose virtual X-ray astrophysics facility that provides access to a carefully selected set of generally useful quantities for individual X-ray sources, and is designed to satisfy the needs of a broad-based group of scientists, including those who may be less familiar with astronomical data analysis in the X-ray regime. The first release of the CSC includes information about 94,676 distinct X-ray sources detected in a subset of public ACIS imaging observations from roughly the first eight years of the Chandra mission. This release of the catalog includes point and compact sources with observed spatial extents <~ 30''. The catalog (1) provides access to the best estimates of the X-ray source properties for detected sources, with good scientific fidelity, and directly supports scientific analysis using the individual source data; (2) facilitates analysis of a wide range of statistical properties for classes of X-ray sources; and (3) provides efficient access to calibrated observational data and ancillary data products for individual X-ray sources, so that users can perform detailed further analysis using existing tools. The catalog includes real X-ray sources detected with flux estimates that are at least 3 times their estimated 1 sigma uncertainties in at least one energy band, while maintaining the number of spurious sources at a level of <~ 1 false source per field for a 100 ks observation. For each detected source, the CSC provides commonly tabulated quantities, including source position, extent, multi-band fluxes, hardness ratios, and variability statistics, derived from the observations in which the source is detected. In addition to these traditional catalog elements, for each X-ray source the CSC includes an extensive set of file-based data products that can be manipulated interactively.
Data downloaded via parachute from a NASA super-pressure balloon
In April to May 2023, the superBIT telescope was lifted to the Earth's stratosphere by a helium-filled super-pressure balloon, to acquire astronomical imaging from above (99.5% of) the Earth's atmosphere. It was launched from New Zealand then, for 40 days, circumnavigated the globe five times at a latitude 40 to 50 degrees South. Attached to the telescope were four 'DRS' (Data Recovery System) capsules containing 5 TB solid state data storage, plus a GNSS receiver, Iridium transmitter, and parachute. Data from the telescope were copied to these, and two were dropped over Argentina. They drifted 61 km horizontally while they descended 32 km, but we predicted their descent vectors within 2.4 km: in this location, the discrepancy appears irreducible below 2 km because of high speed, gusty winds and local topography. The capsules then reported their own locations to within a few metres. We recovered the capsules and successfully retrieved all of superBIT's data - despite the telescope itself being later destroyed on landing.
Overview of the JWST Advanced Deep Extragalactic Survey (JADES)
We present an overview of the James Webb Space Telescope (JWST) Advanced Deep Extragalactic Survey (JADES), an ambitious program of infrared imaging and spectroscopy in the GOODS-S and GOODS-N deep fields, designed to study galaxy evolution from high redshift to cosmic noon. JADES uses about 770 hours of Cycle 1 guaranteed time largely from the Near-Infrared Camera (NIRCam) and Near-Infrared Spectrograph (NIRSpec) instrument teams. In GOODS-S, in and around the Hubble Ultra Deep Field and Chandra Deep Field South, JADES produces a deep imaging region of ~45 arcmin^2 with an average of 130 hrs of exposure time spread over 9 NIRCam filters. This is extended at medium depth in GOODS-S and GOODS-N with NIRCam imaging of ~175 arcmin^2 with an average exposure time of 20 hrs spread over 8-10 filters. In both fields, we conduct extensive NIRSpec multi-object spectroscopy, including 2 deep pointings of 55 hrs exposure time, 14 medium pointings of ~12 hrs, and 15 shallower pointings of ~4 hrs, targeting over 5000 HST and JWST-detected faint sources with 5 low, medium, and high-resolution dispersers covering 0.6-5.3 microns. Finally, JADES extends redward via coordinated parallels with the JWST Mid-Infrared Instrument (MIRI), featuring ~9 arcmin^2 with 43 hours of exposure at 7.7 microns and twice that area with 2-6.5 hours of exposure at 12.8 microns For nearly 30 years, the GOODS-S and GOODS-N fields have been developed as the premier deep fields on the sky; JADES is now providing a compelling start on the JWST legacy in these fields.
Overview of the SDSS-IV MaNGA Survey: Mapping Nearby Galaxies at Apache Point Observatory
We present an overview of a new integral field spectroscopic survey called MaNGA (Mapping Nearby Galaxies at Apache Point Observatory), one of three core programs in the fourth-generation Sloan Digital Sky Survey (SDSS-IV) that began on 2014 July 1. MaNGA will investigate the internal kinematic structure and composition of gas and stars in an unprecedented sample of 10,000 nearby galaxies. We summarize essential characteristics of the instrument and survey design in the context of MaNGA's key science goals and present prototype observations to demonstrate MaNGA's scientific potential. MaNGA employs dithered observations with 17 fiber-bundle integral field units that vary in diameter from 12" (19 fibers) to 32" (127 fibers). Two dual-channel spectrographs provide simultaneous wavelength coverage over 3600-10300 A at R~2000. With a typical integration time of 3 hr, MaNGA reaches a target r-band signal-to-noise ratio of 4-8 (per A, per 2" fiber) at 23 AB mag per sq. arcsec, which is typical for the outskirts of MaNGA galaxies. Targets are selected with stellar mass greater than 1e9 Msun using SDSS-I redshifts and i-band luminosity to achieve uniform radial coverage in terms of the effective radius, an approximately flat distribution in stellar mass, and a sample spanning a wide range of environments. Analysis of our prototype observations demonstrates MaNGA's ability to probe gas ionization, shed light on recent star formation and quenching, enable dynamical modeling, decompose constituent components, and map the composition of stellar populations. MaNGA's spatially resolved spectra will enable an unprecedented study of the astrophysics of nearby galaxies in the coming 6 yr.
Uncovering a Massive z~7.65 Galaxy Hosting a Heavily Obscured Radio-Loud QSO Candidate in COSMOS-Web
In this letter, we report the discovery of the highest redshift, heavily obscured, radio-loud QSO candidate selected using JWST NIRCam/MIRI, mid-IR, sub-mm, and radio imaging in the COSMOS-Web field. Using multi-frequency radio observations and mid-IR photometry, we identify a powerful, radio-loud (RL), growing supermassive black hole (SMBH) with significant spectral steepening of the radio SED (f_{1.32 GHz} sim 2 mJy, q_{24mu m} = -1.1, alpha_{1.32-3GHz}=-1.2, Delta alpha = -0.4). In conjunction with ALMA, deep ground-based observations, ancillary space-based data, and the unprecedented resolution and sensitivity of JWST, we find no evidence of QSO contribution to the UV/optical/NIR data and thus infer heavy amounts of obscuration (N_{H} > 10^{23} cm^{-2}). Using the wealth of deep UV to sub-mm photometric data, we report a singular solution photo-z of z_phot = 7.65^{+0.4}_{-0.3} and estimate an extremely massive host-galaxy (log M_{star} = 11.92 pm 0.06,M_{odot}). This source represents the furthest known obscured RL QSO candidate, and its level of obscuration aligns with the most representative but observationally scarce population of QSOs at these epochs.
Some Theoretical Results on Layerwise Effective Dimension Oscillations in Finite Width ReLU Networks
We analyze the layerwise effective dimension (rank of the feature matrix) in fully-connected ReLU networks of finite width. Specifically, for a fixed batch of m inputs and random Gaussian weights, we derive closed-form expressions for the expected rank of the \mtimes n hidden activation matrices. Our main result shows that E[EDim(ell)]=m[1-(1-2/pi)^ell]+O(e^{-c m}) so that the rank deficit decays geometrically with ratio 1-2 / pi approx 0.3634. We also prove a sub-Gaussian concentration bound, and identify the "revival" depths at which the expected rank attains local maxima. In particular, these peaks occur at depths ell_k^*approx(k+1/2)pi/log(1/rho) with height approx (1-e^{-pi/2}) m approx 0.79m. We further show that this oscillatory rank behavior is a finite-width phenomenon: under orthogonal weight initialization or strong negative-slope leaky-ReLU, the rank remains (nearly) full. These results provide a precise characterization of how random ReLU layers alternately collapse and partially revive the subspace of input variations, adding nuance to prior work on expressivity of deep networks.
Analyzing Leakage of Personally Identifiable Information in Language Models
Language Models (LMs) have been shown to leak information about training data through sentence-level membership inference and reconstruction attacks. Understanding the risk of LMs leaking Personally Identifiable Information (PII) has received less attention, which can be attributed to the false assumption that dataset curation techniques such as scrubbing are sufficient to prevent PII leakage. Scrubbing techniques reduce but do not prevent the risk of PII leakage: in practice scrubbing is imperfect and must balance the trade-off between minimizing disclosure and preserving the utility of the dataset. On the other hand, it is unclear to which extent algorithmic defenses such as differential privacy, designed to guarantee sentence- or user-level privacy, prevent PII disclosure. In this work, we introduce rigorous game-based definitions for three types of PII leakage via black-box extraction, inference, and reconstruction attacks with only API access to an LM. We empirically evaluate the attacks against GPT-2 models fine-tuned with and without defenses in three domains: case law, health care, and e-mails. Our main contributions are (i) novel attacks that can extract up to 10times more PII sequences than existing attacks, (ii) showing that sentence-level differential privacy reduces the risk of PII disclosure but still leaks about 3% of PII sequences, and (iii) a subtle connection between record-level membership inference and PII reconstruction. Code to reproduce all experiments in the paper is available at https://github.com/microsoft/analysing_pii_leakage.
Towards Accurate Generative Models of Video: A New Metric & Challenges
Recent advances in deep generative models have lead to remarkable progress in synthesizing high quality images. Following their successful application in image processing and representation learning, an important next step is to consider videos. Learning generative models of video is a much harder task, requiring a model to capture the temporal dynamics of a scene, in addition to the visual presentation of objects. While recent attempts at formulating generative models of video have had some success, current progress is hampered by (1) the lack of qualitative metrics that consider visual quality, temporal coherence, and diversity of samples, and (2) the wide gap between purely synthetic video data sets and challenging real-world data sets in terms of complexity. To this extent we propose Fr\'{e}chet Video Distance (FVD), a new metric for generative models of video, and StarCraft 2 Videos (SCV), a benchmark of game play from custom starcraft 2 scenarios that challenge the current capabilities of generative models of video. We contribute a large-scale human study, which confirms that FVD correlates well with qualitative human judgment of generated videos, and provide initial benchmark results on SCV.
XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Model
Most Zero-shot Multi-speaker TTS (ZS-TTS) systems support only a single language. Although models like YourTTS, VALL-E X, Mega-TTS 2, and Voicebox explored Multilingual ZS-TTS they are limited to just a few high/medium resource languages, limiting the applications of these models in most of the low/medium resource languages. In this paper, we aim to alleviate this issue by proposing and making publicly available the XTTS system. Our method builds upon the Tortoise model and adds several novel modifications to enable multilingual training, improve voice cloning, and enable faster training and inference. XTTS was trained in 16 languages and achieved state-of-the-art (SOTA) results in most of them.
Efficient Dynamic Clustering-Based Document Compression for Retrieval-Augmented-Generation
Retrieval-Augmented Generation (RAG) has emerged as a widely adopted approach for knowledge integration during large language model (LLM) inference in recent years. However, current RAG implementations face challenges in effectively addressing noise, repetition and redundancy in retrieved content, primarily due to their limited ability to exploit fine-grained inter-document relationships. To address these limitations, we propose an Efficient Dynamic Clustering-based document Compression framework (EDC\textsuperscript{2-RAG}) that effectively utilizes latent inter-document relationships while simultaneously removing irrelevant information and redundant content. We validate our approach, built upon GPT-3.5, on widely used knowledge-QA and hallucination-detected datasets. The results show that this method achieves consistent performance improvements across various scenarios and experimental settings, demonstrating strong robustness and applicability. Our code and datasets can be found at https://github.com/Tsinghua-dhy/EDC-2-RAG.
TransBench: Benchmarking Machine Translation for Industrial-Scale Applications
Machine translation (MT) has become indispensable for cross-border communication in globalized industries like e-commerce, finance, and legal services, with recent advancements in large language models (LLMs) significantly enhancing translation quality. However, applying general-purpose MT models to industrial scenarios reveals critical limitations due to domain-specific terminology, cultural nuances, and stylistic conventions absent in generic benchmarks. Existing evaluation frameworks inadequately assess performance in specialized contexts, creating a gap between academic benchmarks and real-world efficacy. To address this, we propose a three-level translation capability framework: (1) Basic Linguistic Competence, (2) Domain-Specific Proficiency, and (3) Cultural Adaptation, emphasizing the need for holistic evaluation across these dimensions. We introduce TransBench, a benchmark tailored for industrial MT, initially targeting international e-commerce with 17,000 professionally translated sentences spanning 4 main scenarios and 33 language pairs. TransBench integrates traditional metrics (BLEU, TER) with Marco-MOS, a domain-specific evaluation model, and provides guidelines for reproducible benchmark construction. Our contributions include: (1) a structured framework for industrial MT evaluation, (2) the first publicly available benchmark for e-commerce translation, (3) novel metrics probing multi-level translation quality, and (4) open-sourced evaluation tools. This work bridges the evaluation gap, enabling researchers and practitioners to systematically assess and enhance MT systems for industry-specific needs.
XFT: Unlocking the Power of Code Instruction Tuning by Simply Merging Upcycled Mixture-of-Experts
We introduce XFT, a simple yet powerful training scheme, by simply merging upcycled Mixture-of-Experts (MoE) to unleash the performance limit of instruction-tuned code Large Language Models (LLMs). While vanilla sparse upcycling fails to improve instruction tuning, XFT introduces a shared expert mechanism with a novel routing weight normalization strategy into sparse upcycling, which significantly boosts instruction tuning. After fine-tuning the upcycled MoE model, XFT introduces a learnable model merging mechanism to compile the upcycled MoE model back to a dense model, achieving upcycled MoE-level performance with only dense-model compute. By applying XFT to a 1.3B model, we create a new state-of-the-art tiny code LLM (<3B) with 67.1 and 64.6 pass@1 on HumanEval and HumanEval+ respectively. With the same data and model architecture, XFT improves supervised fine-tuning (SFT) by 13% on HumanEval+, along with consistent improvements from 2% to 13% on MBPP+, MultiPL-E, and DS-1000, demonstrating its generalizability. XFT is fully orthogonal to existing techniques such as Evol-Instruct and OSS-Instruct, opening a new dimension for improving code instruction tuning. Codes are available at https://github.com/ise-uiuc/xft .
Sample-Efficient Diffusion for Text-To-Speech Synthesis
This work introduces Sample-Efficient Speech Diffusion (SESD), an algorithm for effective speech synthesis in modest data regimes through latent diffusion. It is based on a novel diffusion architecture, that we call U-Audio Transformer (U-AT), that efficiently scales to long sequences and operates in the latent space of a pre-trained audio autoencoder. Conditioned on character-aware language model representations, SESD achieves impressive results despite training on less than 1k hours of speech - far less than current state-of-the-art systems. In fact, it synthesizes more intelligible speech than the state-of-the-art auto-regressive model, VALL-E, while using less than 2% the training data.
LoRA-Mixer: Coordinate Modular LoRA Experts Through Serial Attention Routing
Recent efforts to combine low-rank adaptation (LoRA) with mixture-of-experts (MoE) for adapting large language models (LLMs) to multiple tasks still exhibit prevailing limitations: they either swap entire attention/feed-forward layers for switch experts or bolt on parallel expert branches, diluting parameter efficiency and task fidelity. We propose the LoRA-Mixer, a modular and lightweight MoE framework that integrates LoRA experts. Our core innovation lies in replacing the projection matrices of the attention module's input/output linear layers with dynamically routed, task-specific LoRA experts. This design ensures seamless compatibility with diverse foundation models, including transformers and state space models (SSMs), by leveraging their inherent linear projection structures. The framework supports two operational paradigms: (1) joint optimization of LoRA experts and routing mechanisms via a novel hard-soft routing strategy, or (2) direct deployment of pre-trained, frozen LoRA modules sourced from external repositories. To enable robust router training with limited data while ensuring stable routing decisions and maximizing expert reuse, we introduce an adaptive Specialization Balance Loss (SBL) that jointly optimizes expert balance and task-specific alignment. Extensive experiments on seven benchmark datasets, including MedQA, CoLA, SST-2, GSM8K, ARC-E, ARC-C, and HumanEval, demonstrate the effectiveness of LoRA-Mixer. On datasets such as GSM8K, HumanEval, and MedQA, LoRA-Mixer achieves significant improvements of 7.61%, 4.88%, and 3.08% over the base models, respectively. Compared with state-of-the-art methods, LoRA-Mixer achieves additional improvements of 1.09%, 1.45%, and 1.68%, respectively, using only 48% of the parameters, demonstrating its efficiency and strong performance.
CalliReader: Contextualizing Chinese Calligraphy via an Embedding-Aligned Vision-Language Model
Chinese calligraphy, a UNESCO Heritage, remains computationally challenging due to visual ambiguity and cultural complexity. Existing AI systems fail to contextualize their intricate scripts, because of limited annotated data and poor visual-semantic alignment. We propose CalliReader, a vision-language model (VLM) that solves the Chinese Calligraphy Contextualization (CC^2) problem through three innovations: (1) character-wise slicing for precise character extraction and sorting, (2) CalliAlign for visual-text token compression and alignment, (3) embedding instruction tuning (e-IT) for improving alignment and addressing data scarcity. We also build CalliBench, the first benchmark for full-page calligraphic contextualization, addressing three critical issues in previous OCR and VQA approaches: fragmented context, shallow reasoning, and hallucination. Extensive experiments including user studies have been conducted to verify our CalliReader's superiority to other state-of-the-art methods and even human professionals in page-level calligraphy recognition and interpretation, achieving higher accuracy while reducing hallucination. Comparisons with reasoning models highlight the importance of accurate recognition as a prerequisite for reliable comprehension. Quantitative analyses validate CalliReader's efficiency; evaluations on document and real-world benchmarks confirm its robust generalization ability.
Comparing Deep Learning Models for Rice Mapping in Bhutan Using High Resolution Satellite Imagery
The Bhutanese government is increasing its utilization of technological approaches such as including Remote Sensing-based knowledge in their decision-making process. This study focuses on crop type and crop extent in Paro, one of the top rice-yielding districts in Bhutan, and employs publicly available NICFI high-resolution satellite imagery from Planet. Two Deep Learning (DL) approaches, point-based (DNN) and patch-based (U-Net), models were used in conjunction with cloud-computing platforms. Three different models per DL approaches (DNN and U-Net) were trained: 1) RGBN channels from Planet; 2) RGBN and elevation data (RGBNE); 3) RGBN and Sentinel-1 (S1) data (RGBNS), and RGBN with E and S1 data (RGBNES). From this comprehensive analysis, the U-Net displayed higher performance metrics across both model training and model validation efforts. Among the U-Net model sets, the RGBN, RGBNE, RGBNS, and RGBNES models had an F1-score of 0.8546, 0.8563, 0.8467, and 0.8500 respectively. An independent model evaluation was performed and found a high level of performance variation across all the metrics. For this independent model evaluation, the U-Net RGBN, RGBNE, RGBNES, and RGBN models displayed the F1-scores of 0.5935, 0.6154, 0.5882, and 0.6582, suggesting U-Net RGBNES as the best model. The study shows that the DL approaches can predict rice. Also, DL methods can be used with the survey-based approaches currently utilized by the Bhutan Department of Agriculture. Further, this study demonstrated the usage of regional land cover products such as SERVIR's RLCMS as a weak label approach to capture different strata addressing the class imbalance problem and improving the sampling design for DL application. Finally, through preliminary model testing and comparisons outlined it was shown that using additional features such as NDVI, EVI, and NDWI did not drastically improve model performance.
Bandit Multi-linear DR-Submodular Maximization and Its Applications on Adversarial Submodular Bandits
We investigate the online bandit learning of the monotone multi-linear DR-submodular functions, designing the algorithm BanditMLSM that attains O(T^{2/3}log T) of (1-1/e)-regret. Then we reduce submodular bandit with partition matroid constraint and bandit sequential monotone maximization to the online bandit learning of the monotone multi-linear DR-submodular functions, attaining O(T^{2/3}log T) of (1-1/e)-regret in both problems, which improve the existing results. To the best of our knowledge, we are the first to give a sublinear regret algorithm for the submodular bandit with partition matroid constraint. A special case of this problem is studied by Streeter et al.(2009). They prove a O(T^{4/5}) (1-1/e)-regret upper bound. For the bandit sequential submodular maximization, the existing work proves an O(T^{2/3}) regret with a suboptimal 1/2 approximation ratio (Niazadeh et al. 2021).
The enigma of the pseudogap phase of the cuprate superconductors
The last few years have seen significant experimental progress in characterizing the copper-based hole-doped high temperature superconductors in the regime of low hole density, p. Quantum oscillations, NMR, X-ray, and STM experiments have shed much light on the nature of the ordering at low temperatures. We review evidence that the order parameter in the non-Lanthanum-based cuprates is a d-form factor density-wave. This novel order acts as an unexpected window into the electronic structure of the pseudogap phase at higher temperatures in zero field: we argue in favor of a `fractionalized Fermi liquid' (FL*) with 4 pockets of spin S=1/2, charge +e fermions enclosing an area specified by p.
CuRe: Cultural Gaps in the Long Tail of Text-to-Image Systems
Popular text-to-image (T2I) systems are trained on web-scraped data, which is heavily Amero and Euro-centric, underrepresenting the cultures of the Global South. To analyze these biases, we introduce CuRe, a novel and scalable benchmarking and scoring suite for cultural representativeness that leverages the marginal utility of attribute specification to T2I systems as a proxy for human judgments. Our CuRe benchmark dataset has a novel categorical hierarchy built from the crowdsourced Wikimedia knowledge graph, with 300 cultural artifacts across 32 cultural subcategories grouped into six broad cultural axes (food, art, fashion, architecture, celebrations, and people). Our dataset's categorical hierarchy enables CuRe scorers to evaluate T2I systems by analyzing their response to increasing the informativeness of text conditioning, enabling fine-grained cultural comparisons. We empirically observe much stronger correlations of our class of scorers to human judgments of perceptual similarity, image-text alignment, and cultural diversity across image encoders (SigLIP 2, AIMV2 and DINOv2), vision-language models (OpenCLIP, SigLIP 2, Gemini 2.0 Flash) and state-of-the-art text-to-image systems, including three variants of Stable Diffusion (1.5, XL, 3.5 Large), FLUX.1 [dev], Ideogram 2.0, and DALL-E 3. The code and dataset is open-sourced and available at https://aniketrege.github.io/cure/.
E2 TTS: Embarrassingly Easy Fully Non-Autoregressive Zero-Shot TTS
This paper introduces Embarrassingly Easy Text-to-Speech (E2 TTS), a fully non-autoregressive zero-shot text-to-speech system that offers human-level naturalness and state-of-the-art speaker similarity and intelligibility. In the E2 TTS framework, the text input is converted into a character sequence with filler tokens. The flow-matching-based mel spectrogram generator is then trained based on the audio infilling task. Unlike many previous works, it does not require additional components (e.g., duration model, grapheme-to-phoneme) or complex techniques (e.g., monotonic alignment search). Despite its simplicity, E2 TTS achieves state-of-the-art zero-shot TTS capabilities that are comparable to or surpass previous works, including Voicebox and NaturalSpeech 3. The simplicity of E2 TTS also allows for flexibility in the input representation. We propose several variants of E2 TTS to improve usability during inference. See https://aka.ms/e2tts/ for demo samples.
LooseControl: Lifting ControlNet for Generalized Depth Conditioning
We present LooseControl to allow generalized depth conditioning for diffusion-based image generation. ControlNet, the SOTA for depth-conditioned image generation, produces remarkable results but relies on having access to detailed depth maps for guidance. Creating such exact depth maps, in many scenarios, is challenging. This paper introduces a generalized version of depth conditioning that enables many new content-creation workflows. Specifically, we allow (C1) scene boundary control for loosely specifying scenes with only boundary conditions, and (C2) 3D box control for specifying layout locations of the target objects rather than the exact shape and appearance of the objects. Using LooseControl, along with text guidance, users can create complex environments (e.g., rooms, street views, etc.) by specifying only scene boundaries and locations of primary objects. Further, we provide two editing mechanisms to refine the results: (E1) 3D box editing enables the user to refine images by changing, adding, or removing boxes while freezing the style of the image. This yields minimal changes apart from changes induced by the edited boxes. (E2) Attribute editing proposes possible editing directions to change one particular aspect of the scene, such as the overall object density or a particular object. Extensive tests and comparisons with baselines demonstrate the generality of our method. We believe that LooseControl can become an important design tool for easily creating complex environments and be extended to other forms of guidance channels. Code and more information are available at https://shariqfarooq123.github.io/loose-control/ .
Versatile Diffusion: Text, Images and Variations All in One Diffusion Model
The recent advances in diffusion models have set an impressive milestone in many generation tasks. Trending works such as DALL-E2, Imagen, and Stable Diffusion have attracted great interest in academia and industry. Despite the rapid landscape changes, recent new approaches focus on extensions and performance rather than capacity, thus requiring separate models for separate tasks. In this work, we expand the existing single-flow diffusion pipeline into a multi-flow network, dubbed Versatile Diffusion (VD), that handles text-to-image, image-to-text, image-variation, and text-variation in one unified model. Moreover, we generalize VD to a unified multi-flow multimodal diffusion framework with grouped layers, swappable streams, and other propositions that can process modalities beyond images and text. Through our experiments, we demonstrate that VD and its underlying framework have the following merits: a) VD handles all subtasks with competitive quality; b) VD initiates novel extensions and applications such as disentanglement of style and semantic, image-text dual-guided generation, etc.; c) Through these experiments and applications, VD provides more semantic insights of the generated outputs. Our code and models are open-sourced at https://github.com/SHI-Labs/Versatile-Diffusion.
F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching
This paper introduces F5-TTS, a fully non-autoregressive text-to-speech system based on flow matching with Diffusion Transformer (DiT). Without requiring complex designs such as duration model, text encoder, and phoneme alignment, the text input is simply padded with filler tokens to the same length as input speech, and then the denoising is performed for speech generation, which was originally proved feasible by E2 TTS. However, the original design of E2 TTS makes it hard to follow due to its slow convergence and low robustness. To address these issues, we first model the input with ConvNeXt to refine the text representation, making it easy to align with the speech. We further propose an inference-time Sway Sampling strategy, which significantly improves our model's performance and efficiency. This sampling strategy for flow step can be easily applied to existing flow matching based models without retraining. Our design allows faster training and achieves an inference RTF of 0.15, which is greatly improved compared to state-of-the-art diffusion-based TTS models. Trained on a public 100K hours multilingual dataset, our Fairytaler Fakes Fluent and Faithful speech with Flow matching (F5-TTS) exhibits highly natural and expressive zero-shot ability, seamless code-switching capability, and speed control efficiency. Demo samples can be found at https://SWivid.github.io/F5-TTS. We release all code and checkpoints to promote community development.
Governance of the AI, by the AI, and for the AI
Over the past half century, there have been several false dawns during which the "arrival" of world-changing artificial intelligence (AI) has been heralded. Tempting fate, the authors believe the age of AI has, indeed, finally arrived. Powerful image generators, such as DALL-E2 and Midjourney have suddenly allowed anyone with access the ability easily to create rich and complex art. In a similar vein, text generators, such as GPT3.5 (including ChatGPT) and BLOOM, allow users to compose detailed written descriptions of many topics of interest. And, it is even possible now for a person without extensive expertise in writing software to use AI to generate code capable of myriad applications. While AI will continue to evolve and improve, probably at a rapid rate, the current state of AI is already ushering in profound changes to many different sectors of society. Every new technology challenges the ability of humanity to govern it wisely. However, governance is usually viewed as both possible and necessary due to the disruption new technology often poses to social structures, industries, the environment, and other important human concerns. In this article, we offer an analysis of a range of interactions between AI and governance, with the hope that wise decisions may be made that maximize benefits and minimize costs. The article addresses two main aspects of this relationship: the governance of AI by humanity, and the governance of humanity by AI. The approach we have taken is itself informed by AI, as this article was written collaboratively by the authors and ChatGPT.
The Carbon Emissions of Writing and Illustrating Are Lower for AI than for Humans
As AI systems proliferate, their greenhouse gas emissions are an increasingly important concern for human societies. We analyze the emissions of several AI systems (ChatGPT, BLOOM, DALL-E2, Midjourney) relative to those of humans completing the same tasks. We find that an AI writing a page of text emits 130 to 1500 times less CO2e than a human doing so. Similarly, an AI creating an image emits 310 to 2900 times less. Emissions analysis do not account for social impacts such as professional displacement, legality, and rebound effects. In addition, AI is not a substitute for all human tasks. Nevertheless, at present, the use of AI holds the potential to carry out several major activities at much lower emission levels than can humans.
Expanding Small-Scale Datasets with Guided Imagination
The power of DNNs relies heavily on the quantity and quality of training data. However, collecting and annotating data on a large scale is often expensive and time-consuming. To address this issue, we explore a new task, termed dataset expansion, aimed at expanding a ready-to-use small dataset by automatically creating new labeled samples. To this end, we present a Guided Imagination Framework (GIF) that leverages cutting-edge generative models like DALL-E2 and Stable Diffusion (SD) to "imagine" and create informative new data from the input seed data. Specifically, GIF conducts data imagination by optimizing the latent features of the seed data in the semantically meaningful space of the prior model, resulting in the creation of photo-realistic images with new content. To guide the imagination towards creating informative samples for model training, we introduce two key criteria, i.e., class-maintained information boosting and sample diversity promotion. These criteria are verified to be essential for effective dataset expansion: GIF-SD obtains 13.5% higher model accuracy on natural image datasets than unguided expansion with SD. With these essential criteria, GIF successfully expands small datasets in various scenarios, boosting model accuracy by 36.9% on average over six natural image datasets and by 13.5% on average over three medical datasets. The source code is available at https://github.com/Vanint/DatasetExpansion.
Pseudo-Autoregressive Neural Codec Language Models for Efficient Zero-Shot Text-to-Speech Synthesis
Recent zero-shot text-to-speech (TTS) systems face a common dilemma: autoregressive (AR) models suffer from slow generation and lack duration controllability, while non-autoregressive (NAR) models lack temporal modeling and typically require complex designs. In this paper, we introduce a novel pseudo-autoregressive (PAR) codec language modeling approach that unifies AR and NAR modeling. Combining explicit temporal modeling from AR with parallel generation from NAR, PAR generates dynamic-length spans at fixed time steps. Building on PAR, we propose PALLE, a two-stage TTS system that leverages PAR for initial generation followed by NAR refinement. In the first stage, PAR progressively generates speech tokens along the time dimension, with each step predicting all positions in parallel but only retaining the left-most span. In the second stage, low-confidence tokens are iteratively refined in parallel, leveraging the global contextual information. Experiments demonstrate that PALLE, trained on LibriTTS, outperforms state-of-the-art systems trained on large-scale data, including F5-TTS, E2-TTS, and MaskGCT, on the LibriSpeech test-clean set in terms of speech quality, speaker similarity, and intelligibility, while achieving up to ten times faster inference speed. Audio samples are available at https://anonymous-palle.github.io.