Archive

Find an article within this site using search terms:

All Blog Posts in Reverse Chronological Order

• 11 Apr 2025 » Defending against Prompt Injection with Structured Queries (StruQ) and Preference Optimization (SecAlign)

• 08 Apr 2025 » Repurposing Protein Folding Models for Generation with Latent Diffusion

• 25 Mar 2025 » Scaling Up Reinforcement Learning for Traffic Smoothing: A 100-AV Highway Deployment

• 12 Nov 2024 » Virtual Personas for Language Models via an Anthology of Backstories

• 20 Sep 2024 » Linguistic Bias in ChatGPT: Language Models Reinforce Dialect Discrimination

• 28 Aug 2024 » How to Evaluate Jailbreak Methods: A Case Study with the StrongREJECT Benchmark

• 20 Jul 2024 » Are We Ready for Multi-Image Reasoning? Launching VHs: The Visual Haystacks Benchmark!

• 29 May 2024 » TinyAgent: Function Calling at the Edge

• 21 Mar 2024 » Modeling Extremely Large Images with xT

• 11 Mar 2024 » 2024 BAIR Graduate Directory

• 18 Feb 2024 » The Shift from Models to Compound AI Systems

• 14 Nov 2023 » Ghostbuster: Detecting Text Ghostwritten by Large Language Models

• 14 Nov 2023 » Asymmetric Certified Robustness via Feature-Convex Neural Networks

• 17 Oct 2023 » Goal Representations for Instruction Following

• 16 Oct 2023 » Rethinking the Role of PPO in RLHF

• 14 Jul 2023 » Training Diffusion Models with
  Reinforcement Learning

• 10 Jul 2023 » On the Stepwise Nature of
  Self-Supervised Learning

• 29 Jun 2023 » Generating 3D Molecular Conformers via Equivariant Coarse-Graining and Aggregated Attention

• 23 May 2023 » GPT-4 + Stable-Diffusion = ?: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models

• 06 Apr 2023 » Interactive Fleet Learning

• 03 Apr 2023 » Koala: A Dialogue Model for Academic Research

• 20 Jan 2023 » Fully Autonomous Real-World Reinforcement Learning with Applications to Mobile Manipulation

• 19 Sep 2022 » Keeping Learning-Based Control Safe by Regulating Distributional Shift

• 29 Aug 2022 » Reverse engineering the NTK: towards first-principles architecture design

• 10 Jul 2022 » Why do Policy Gradient Methods work so well in Cooperative MARL? Evidence from Policy Representation

• 30 Jun 2022 » FIGS: Attaining XGBoost-level performance with the interpretability and speed of CART

• 20 May 2022 » The Berkeley Crossword Solver

• 03 May 2022 » Rethinking Human-in-the-Loop for Artificial Augmented Intelligence

• 29 Apr 2022 » Designing Societally Beneficial Reinforcement Learning Systems

• 21 Apr 2022 » Should I Use Offline RL or Imitation Learning?

• 20 Apr 2022 » Offline RL Made Easier: No TD Learning, Advantage Reweighting, or Transformers

• 21 Mar 2022 » Accelerating Ukraine Intelligence Analysis with Computer Vision on Synthetic Aperture Radar Imagery

• 23 Feb 2022 » Unsupervised Skill Discovery with Contrastive Intrinsic Control

• 02 Feb 2022 » imodels: leveraging the unreasonable effectiveness of rules

• 14 Dec 2021 » The Unsupervised Reinforcement Learning Benchmark

• 19 Nov 2021 » Sequence Modeling Solutions
  for Reinforcement Learning Problems

• 19 Nov 2021 » Which Mutual Information Representation Learning Objectives are Sufficient for Control?

• 17 Nov 2021 » Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasets

• 08 Nov 2021 » How should we compare neural network representations?

• 03 Nov 2021 » RECON: Learning to Explore the Real World with a Ground Robot

• 01 Nov 2021 » Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability

• 25 Oct 2021 » Designs from Data: Offline Black-Box Optimization via Conservative Training

• 25 Oct 2021 » A First-Principles Theory of Neural
  Network Generalization

• 22 Oct 2021 » Making RL Tractable by Learning More Informative Reward Functions: Example-Based Control, Meta-Learning, and Normalized Maximum Likelihood

• 14 Oct 2021 » Updates and Lessons from AI Forecasting

• 06 Oct 2021 » PICO: Pragmatic Compression for Human-in-the-Loop Decision-Making

• 29 Sep 2021 » Unsolved ML Safety Problems

• 28 Sep 2021 » Distilling neural networks into wavelet models using interpretations

• 24 Sep 2021 » What Can I Do Here? Learning New Skills by Imagining Visual Affordances

• 22 Jul 2021 » Universal Weakly Supervised Segmentation by Pixel-to-Segment Contrastive Learning

• 14 Jul 2021 » The Surprising Effectiveness of PPO in Cooperative Multi-Agent Games

• 08 Jul 2021 » BASALT: A Benchmark for
  Learning from Human Feedback

• 03 May 2021 » Learning What To Do by Simulating the Past

• 20 Apr 2021 » An EPIC way to evaluate reward functions

• 19 Apr 2021 » The Importance of Hyperparameter Optimization for Model-based Reinforcement Learning

• 23 Mar 2021 » Pretrained Transformers as Universal Computation Engines

• 09 Mar 2021 » Maximum Entropy RL (Provably) Solves Some Robust RL Problems

• 25 Feb 2021 » Self-Supervised Policy Adaptation during Deployment

• 05 Jan 2021 » The Successor Representation, $\gamma$-Models,
  and Infinite-Horizon Prediction

• 20 Dec 2020 » Does GPT-2 Know Your Phone Number?

• 07 Dec 2020 » Offline Reinforcement Learning: How Conservative Algorithms Can Enable New Applications

• 20 Nov 2020 » Learning State Abstractions for Long-Horizon Planning

• 18 Nov 2020 » EvolveGraph: Dynamic Neural Relational Reasoning for Interacting Systems

• 16 Nov 2020 » Training on Test Inputs with Amortized Conditional Normalized Maximum Likelihood

• 13 Nov 2020 » Goodhart’s Law, Diversity and a Series of Seemingly Unrelated Toy Problems

• 05 Nov 2020 » Adapting on the Fly to Test Time Distribution Shift

• 13 Oct 2020 » Reinforcement learning is supervised learning on optimized data

• 06 Oct 2020 » Plan2Explore: Active Model-Building for Self-Supervised Visual Reinforcement Learning

• 10 Sep 2020 » AWAC: Accelerating Online Reinforcement Learning with Offline Datasets

• 16 Aug 2020 » AI Will Change the World.
  Who Will Change AI?
  We Will.

• 03 Aug 2020 » Estimating the fatality rate is difficult but doable with better data

• 24 Jul 2020 » Exploring Exploration: Comparing Children with RL Agents in Unified Environments

• 19 Jul 2020 » Can RL From Pixels be as Efficient as RL From State?

• 11 Jul 2020 » Decentralized Reinforcement Learning:
  Global Decision-Making via
  Local Economic Transactions

• 25 Jun 2020 » D4RL: Building Better Benchmarks for Offline Reinforcement Learning

• 14 Jun 2020 » Open Compound Domain Adaptation

• 14 May 2020 » OmniTact: A Multi-Directional High-Resolution Touch Sensor

• 05 May 2020 » Four Novel Approaches to Manipulating Fabric using Model-Free and Model-Based Deep Learning in Simulation

• 01 May 2020 » Unsupervised Meta-Learning: Learning to Learn without Supervision

• 27 Apr 2020 » The Ingredients of Real World Robotic Reinforcement Learning

• 23 Apr 2020 » Making Decision Trees Accurate Again: Explaining What Explainable AI Did Not

• 03 Apr 2020 » Robots Learning to Move like Animals

• 27 Mar 2020 » Physically Realistic Attacks on Deep Reinforcement Learning

• 16 Mar 2020 » Does On-Policy Data Collection Fix Errors in Off-Policy Reinforcement Learning?

• 12 Mar 2020 » BADGR:
  The Berkeley Autonomous Driving Ground Robot

• 05 Mar 2020 » Speeding Up Transformer Training and Inference By Increasing Model Size

• 16 Jan 2020 » Large Scale Training at BAIR with Ray Tune

• 18 Dec 2019 » Emergent Behavior by Minimizing Chaos

• 16 Dec 2019 » What is My Data Worth?

• 13 Dec 2019 » Learning to Imitate Human Demonstrations via CycleGAN

• 12 Dec 2019 » Model-Based Reinforcement Learning:
  Theory and Practice

• 05 Dec 2019 » Data-Driven Deep Reinforcement Learning

• 26 Nov 2019 » RoboNet: A Dataset for Large-Scale Multi-Robot Learning

• 22 Nov 2019 » Prof. Anca Dragan Talks About Human-Robot Interaction for WIRED

• 04 Nov 2019 » Can We Learn the Language of Proteins?

• 28 Oct 2019 » Look then Listen: Pre-Learning Environment Representations for Data-Efficient Neural Instruction Following

• 21 Oct 2019 » Collaborating with Humans Requires Understanding Them

• 14 Oct 2019 » Functional RL with Keras and Tensorflow Eager

• 30 Sep 2019 » Deep Dynamics Models for Dexterous Manipulation

• 26 Sep 2019 » Sample Efficient Evolutionary Algorithm for Analog Circuit Design

• 24 Sep 2019 » rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorch

• 19 Sep 2019 » A Deep Learning Approach to Data Compression

• 13 Aug 2019 » Evaluating and Testing Unintended Memorization in Neural Networks

• 10 Jun 2019 » Learning to Learn with Probabilistic Task Embeddings

• 07 Jun 2019 » 1000x Faster Data Augmentation

• 03 Jun 2019 » Autonomous Vehicles for Social Good: Learning to Solve Congestion

• 28 May 2019 » End-to-End Deep Reinforcement Learning
  without Reward Engineering

• 20 May 2019 » Model-Based Reinforcement Learning from Pixels with Structured Latent Variable Models

• 13 May 2019 » Large-Scale Long-Tailed Recognition in an Open World

• 06 May 2019 » Robots that Learn to Adapt

• 11 Apr 2019 » Robots that Learn to Use Improvised Tools

• 27 Mar 2019 » CVPR 2019 Challenges on Domain Adaptation in Autonomous Driving

• 24 Mar 2019 » Announcing the BAIR Open Research Commons

• 21 Mar 2019 » Manipulation By Feel

• 18 Mar 2019 » Assessing Generalization in Deep Reinforcement Learning

• 15 Feb 2019 » Controlling False Discoveries in Large-Scale Experimentation: Challenges and Solutions

• 11 Feb 2019 » Learning Preferences by Looking at the World

• 14 Dec 2018 » Soft Actor Critic—Deep Reinforcement Learning with Real-World Robots

• 12 Dec 2018 » Scaling Multi-Agent Reinforcement Learning

• 05 Dec 2018 » Building Gene Expression Atlases with Deep Generative Models for Single-cell Transcriptomics

• 30 Nov 2018 » Visual Model-Based Reinforcement Learning as a Path towards Generalist Robots

• 26 Nov 2018 » Physics-Based Learned Design: Teaching a Microscope How to Image

• 14 Nov 2018 » AdaSearch: A Successive Elimination Approach to Adaptive Search

• 23 Oct 2018 » Drilling Down on Depth Sensing and Deep Learning

• 09 Oct 2018 » Learning Acrobatics by Watching YouTube

• 06 Sep 2018 » Visual Reinforcement Learning with Imagined Goals

• 31 Aug 2018 » Dexterous Manipulation with Reinforcement Learning: Efficient, General, and Low-Cost

• 06 Aug 2018 » When Recurrent Models Don’t Need to be Recurrent

• 28 Jun 2018 » One-Shot Imitation from Watching Videos

• 18 Jun 2018 » BDD100K Blog Update

• 30 May 2018 » BDD100K: A Large-scale Diverse Driving Video Database

• 17 May 2018 » Delayed Impact of Fair Machine Learning

• 26 Apr 2018 » TDM: From Model-Free to Model-Based Deep Reinforcement Learning

• 18 Apr 2018 » Shared Autonomy via Deep Reinforcement Learning

• 10 Apr 2018 » Towards a Virtual Stuntman

• 13 Mar 2018 » Transfer Your Font Style with GANs

• 06 Feb 2018 » Learning Robot Objectives from Physical Human Interaction

• 23 Jan 2018 » Kernel Feature Selection via Conditional Covariance Minimization

• 09 Jan 2018 » Ray: A Distributed System for AI

• 30 Dec 2017 » Physical Adversarial Examples Against Deep Neural Networks

• 20 Dec 2017 » Reverse Curriculum Generation for Reinforcement Learning Agents

• 12 Dec 2017 » Towards Intelligent Industrial Co-robots

• 05 Dec 2017 » FaSTrack: Ensuring Safe Real-Time Navigation of Dynamic Systems

• 30 Nov 2017 » Model-based Reinforcement Learning with Neural Network Dynamics

• 09 Nov 2017 » The Emergence of a Fovea while Learning to Attend

• 26 Oct 2017 » DART: Noise Injection for Robust Imitation Learning

• 17 Oct 2017 » Learning Long Duration Sequential Task Structure From Demonstrations with Application in Surgical Robotics

• 06 Oct 2017 » Learning Diverse Skills via Maximum Entropy Deep Reinforcement Learning

• 12 Sep 2017 » Learning to Optimize with Reinforcement Learning

• 05 Sep 2017 » Learning a Multi-View Stereo Machine

• 31 Aug 2017 » How to Escape Saddle Points Efficiently

• 23 Aug 2017 » High Quality 3D Object Reconstruction from a Single Color Image

• 17 Aug 2017 » Cooperatively Learning Human Values

• 08 Aug 2017 » Captioning Novel Objects in Images

• 02 Aug 2017 » Minibatch Metropolis-Hastings

• 18 Jul 2017 » Learning to Learn

• 11 Jul 2017 » The Confluence of Geometry and Learning

• 06 Jul 2017 » Constrained Policy Optimization

• 27 Jun 2017 » Releasing the Dexterity Network (Dex-Net) 2.0 Dataset for Deep Grasping

• 20 Jun 2017 » Learning to Reason with Neural Module Networks

• 20 Jun 2017 » Introducing the BAIR Blog