1

Discovering state-of-the-art reinforcement learning algorithms

DataRater: Meta-Learned Dataset Curation

Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities

Learning from negative feedback, or positive feedback or both

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

Gemini: A Family of Highly Capable Multimodal Models

Deep Reinforcement Learning with Plasticity Injection

In-context Reinforcement Learning with Algorithm Distillation

Introducing Symmetries to Black Box Meta Reinforcement Learning

Discovery of Options via Meta-Learned Subgoals