Research
Publications
-
Research
Our latest advances in robot dexterity
Two new AI systems, ALOHA Unleashed and DemoStart, help robots learn to perform complex tasks that require dexterous movement
-
Research
AlphaProteo generates novel proteins for biology and health research
New AI system designs proteins that successfully bind to target molecules, with potential for advancing drug design, disease understanding and more.
-
Research
FermiNet: Quantum physics and chemistry from first principles
Using deep learning to solve fundamental problems in computational quantum chemistry and explore how matter interacts with light
Publications
Explore a selection of our recent research on some of the most complex and interesting challenges in AI.
-
- Date
- Title
- Privacy Amplification by Sampling for the Matrix Mechanism.
- Authors
- Venue
-
- Date
- Title
- From Sparse to Soft Mixture of Experts
- Authors
- Venue
-
- Date
- Title
- π2vec: Policy Representations with SuccessorFeatures
- Authors
- Venue
-
- Date
- Title
- Language Modeling Is Compression
- Authors
- Venue
-
- Date
- Title
- ExeDec: Execution Decomposition for Compositional Generalization in Neural Program Synthesis
- Authors
- Venue
-
- Date
- Title
- Position: Leverage Foundational Models for Black-Box Optimization
- Authors
- Venue
-
- Date
- Title
- Improving Dictionary Learning with Gated Sparse Autoencoders
- Authors
- Venue
-
- Date
- Title
- Holistic Safety and Responsibility Evaluations of Advanced AI Models
- Authors
- Venue
-
- Date
- Title
- Many-Shot In-Context Learning
- Authors
- Venue
-
- Date
- Title
- Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning
- Authors
- Venue
-
- Date
- Title
- Gecko: Versatile Text Embeddings Distilled from Large Language Models
- Authors
- Venue
-
- Date
- Title
- Learning from One Continuous Video Stream
- Authors
- Venue
-
- Date
- Title
- Few-Shot Recalibration of Language Models
- Authors
- Venue
-
- Date
- Title
- Long-form factuality in large language models
- Authors
- Venue
-
- Date
- Title
- Evaluating Frontier Models for Dangerous Capabilities
- Authors
- Venue
-
- Date
- Title
- DiPaCo: Distributed Path Composition
- Authors
- Venue
-
- Date
- Title
- Robust Exploration via Clustering-based Density Estimation
- Authors
- Venue
-
- Date
- Title
- Model-free Posterior Sampling via Learning Rate Randomization
- Authors
- Venue
-
- Date
- Title
- Understanding Learning from Human Preferences
- Authors
- Venue