Ali Razavi
Papers
1
Total Citations
66
H-Index
1
About
Ali Razavi is a leading researcher in artificial intelligence, whose work sits at the intersection of large-scale modeling, multi-modal learning, and generalist agent design. Razavi’s most significant contribution is the development of Gato, a groundbreaking generalist agent introduced in the highly cited 2022 paper "A Generalist Agent" (66 citations). Inspired by the scaling successes of large language models, Razavi and the team built a single, unified neural network capable of performing hundreds of diverse tasks—from playing Atari games to stacking blocks with a real robot arm—using the same set of weights. This work represents a paradigm shift from specialized AI to a more flexible, multi-task, and multi-embodiment policy, demonstrating that a single agent can learn to act across vastly different environments and output modalities. By showing that a transformer-based architecture can serve as a generalist policy, Razavi’s research has opened new avenues for building more adaptable and capable AI systems, directly influencing the trajectory of foundation models for embodied intelligence.
Research Focus
Key Achievements
Top Papers
- 1A Generalist Agent66 citations · 2022