Using geometric problem solving as a surrogate task to enhance models' spatial intelligence capabilities.
Shijie Lian
LiamLian0727
AI & ML interests
VLM and VLA
Recent Activity
commented on
a paper
about 15 hours ago
BayesianVLA: Bayesian Decomposition of Vision Language Action Models via Latent Action Queries
upvoted
a
paper
1 day ago
VideoMaMa: Mask-Guided Video Matting via Generative Prior
submitted
a paper
2 days ago
BayesianVLA: Bayesian Decomposition of Vision Language Action Models via Latent Action Queries