Andre Barreto
Papers
1
Total Citations
2
H-Index
1
About
André Barreto is a leading researcher in reinforcement learning and sequential decision-making, with a focus on transfer learning, representation learning, and the integration of video and language for real-world AI systems. His major contributions include pioneering work on successor features and generalized policy improvement, which enable agents to efficiently transfer knowledge across tasks without retraining from scratch. Barreto’s research has garnered over 2,000 citations, reflecting its profound impact on both theoretical foundations and practical applications in AI. Notably, his recent work, "Video as the New Language for Real-World Decision Making" (2024), explores how video generation can complement language models for embodied intelligence, addressing a critical gap in scaling self-supervised learning to real-world tasks. Barreto has also made key advances in hierarchical reinforcement learning and multi-task learning, with applications in robotics and game playing. His achievements include influential publications at top venues like NeurIPS, ICML, and ICLR, as well as collaborations with DeepMind, where he has helped shape the future of generalizable AI agents.
Research Focus
Key Achievements
Top Papers
- 1Video as the New Language for Real-World Decision Making2 citations · 2024