Google DeepMind

Research AI system associated with protein-structure prediction and scientific discovery.

Google DeepMind model for predicting the effects of genetic variants on genome function.

DeepMind system that combined neural networks and search to achieve superhuman Go performance.

Google DeepMind model for predicting the pathogenicity of missense genetic variants.

DeepMind reinforcement-learning system developed to play StarCraft II at grandmaster level.

General reinforcement-learning system that learned chess, shogi and Go from self-play.

DeepMind generalist agent research model trained across language, vision, games and robot-control tasks.

Google DeepMind's family of vision-language-action and embodied-reasoning models for robotics.

Google DeepMind probabilistic weather model designed to generate ensembles of future weather scenarios.

Google DeepMind world-model research for generating interactive environments.

Google DeepMind machine-learning weather forecasting system based on graph neural networks.

Google DeepMind's text-to-image model family, including Imagen 4.

Google DeepMind's generative video model family.

DeepMind generative neural network for raw audio waveforms that influenced neural speech synthesis.