Everything you care about in one place

Follow feeds: blogs, news, RSS and more. An effortless way to read and digest content of your choice.

Get Feeder

bair.berkeley.edu

The Berkeley Artificial Intelligence Research Blog

Get the latest updates from The Berkeley Artificial Intelligence Research Blog directly as they happen.

Follow now 382 followers

Latest posts

Last updated about 11 hours ago

Teaching LLMs to Update Beliefs for Efficient Long-Horizon Interaction

1 day ago

Overview of ABBEL compared to traditional recursive summarization. Beliefs replace the full...

Intelligence is Free, Now What? Data Systems for, of, and by Agents

20 days ago

... government of the people, by the people, for the people ......

2026 BAIR Graduate Showcase

26 days ago

Congratulations to the Berkeley Artificial Intelligence Research (BAIR) Lab class of 2026!...

Adaptive Parallel Reasoning: The Next Paradigm in Efficient Inference Scaling

3 months ago

Overview of adaptive parallel reasoning. What if a reasoning model could decide...

Gradient-based Planning for World Models at Longer Horizons

3 months ago

GRASP is a new gradient-based planner for learned dynamics (a “world model”)...

Information-Driven Design of Imaging Systems

7 months ago

An encoder (optical system) maps objects to noiseless images, which noise corrupts...

Identifying Interactions at Scale for LLMs

5 months ago

Understanding the behavior of complex machine learning systems, particularly Large Language Models...

Whole-Body Conditioned Egocentric Video Prediction

about 1 year ago

× Predicting Ego-centric Video from human Actions (PEVA). Given past video frames...

What exactly does word2vec learn?

11 months ago

What exactly does word2vec learn, and how? Answering this question amounts to...

RL without TD learning

9 months ago

In this post, I’ll introduce a reinforcement learning (RL) algorithm based on...

Defending against Prompt Injection with Structured Queries (StruQ) and Preference Optimization (SecAlign)

over 1 year ago

Recent advances in Large Language Models (LLMs) enable exciting LLM-integrated applications. However,...

Repurposing Protein Folding Models for Generation with Latent Diffusion

over 1 year ago

PLAID is a multimodal generative model that simultaneously generates protein 1D sequence...