Papers
12
Total Citations
276
H-Index
8
About
Jake Bruce is an artificial intelligence researcher whose work spans robotics, reinforcement learning, and embodied AI. He is perhaps best known as a co-author of "A Generalist Agent" (2022, 66 citations), which introduced Gato — DeepMind's landmark multi-modal, multi-task generalist policy capable of performing hundreds of diverse tasks within a single network, representing a significant milestone in the pursuit of general-purpose AI systems. His broader research consistently tackles one of robotics' most persistent challenges: enabling agents to learn effective navigation and interaction policies without prohibitive real-world data collection, as demonstrated in his work on one-shot reinforcement learning and kilometer-scale deployable navigation policies. Bruce also made early contributions to vision-and-language navigation, helping bridge natural language instruction-following with real-world robotic environments (62 citations). His human-robot interaction research — including gesture-based UAV control and robust multimodal sensor fusion in crowds — reflects a commitment to making autonomous systems genuinely responsive to uninstrumented humans. Additional work on training stability through "Ctrl-Z" and sim-to-real transfer underscores his attention to the practical safety and reliability demands of deploying learned policies beyond the laboratory.
Research Focus
Key Achievements
Top Papers
- 1A Generalist Agent66 citations · 2022
- 2
- 3One-Shot Reinforcement Learning for Robot Navigation with Interactive Replay53 citations · 2017
- 4
- 5
- 6Robust sensor fusion for finding HRI partners in a crowd13 citations · 2017
- 7Ctrl-Z: Recovering from Instability in Reinforcement Learning13 citations · 2019
- 8Ready—Aim—Fly! Hands-Free Face-Based HRI for 3D Trajectory Control of UAVs11 citations · 2017
- 9Zero-shot Sim-to-Real Transfer with Modular Priors.2 citations · 2018
- 10