Diffusion models (DDPM), NeRF, robot learning, and open chat models such as Vicuna.
Landmark work
Papers on Papers
Lab page on PapersCorrection Training for Diffusion Models
OmniTaskonomy: When Does Visual Generation Improve Visual Understanding?
FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution
Playful Agentic Robot Learning
Agents' Last Exam
Flash-KMeans: Fast and Memory-Efficient Exact K-Means
dLLM: Simple Diffusion Language Modeling
SLA2: Sparse-Linear Attention with Learnable Routing and QAT
Links
- Website
- bair.berkeley.edu
- Blog
- bair.berkeley.edu/blog
- X
- @berkeley_ai
- Papers
- Lab page on Papers
About, sources and history
- Type
- Academic, ai research
- Part of
- UC Berkeley
- Sources
- bair.berkeley.edu
Related labs: Berkeley Sky Computing Lab, Center for Human-Compatible Artificial Intelligence (CHAI)
History