Richard Sutton
One of the founding researchers of reinforcement learning, co-author of Reinforcement Learning: An Introduction with Andrew Barto, co-founder of Keen Technologies alongside John Carmack, and recipient of the 2024 Turing Award.
Core Facts
Timeline (last 90 days)
The Bitter Lesson proposed by Richard Sutton points out that in the long run, AI systems that leverage large-scale computation and general learning methods such as search and learning will ultimately prevail over those systems that rely on human domain knowledge and hand-designed features.
The Dyna architecture was proposed by reinforcement learning pioneer Richard Sutton.
The core argument of The Bitter Lesson is that methods relying on human knowledge, domain priors, and hand-crafted rules perform well in the short term, but in the long run are always surpassed by general methods that rely solely on compute and data.
Dyna-Q was proposed by reinforcement learning pioneer Richard Sutton in 1991 and is a method that combines model-based learning with model-free learning.
《Reinforcement Learning: An Introduction》由Richard Sutton和Andrew Barto合著，第二版于2018年更新，全书可免费在线阅读
时序差分（TD）学习由 Richard Sutton 在1988年正式提出，结合了蒙特卡洛方法和动态规划的优点
All Facts (8)
Richard Sutton and Andrew Barto co-authored "Reinforcement Learning: An Introduction," systematically laying the theoretical framework for the field of reinforcement learning.
65%UnverifiedRichard Sutton and Andrew Barto systematically formalized the RL theoretical framework in their 1998 book Reinforcement Learning: An Introduction
60%UnverifiedThe Bitter Lesson proposed by Richard Sutton points out that in the long run, AI systems that leverage large-scale computation and general learning methods such as search and learning will ultimately prevail over those systems that rely on human domain knowledge and hand-designed features.
50%UnverifiedThe Dyna architecture was proposed by reinforcement learning pioneer Richard Sutton.
50%UnverifiedThe core argument of The Bitter Lesson is that methods relying on human knowledge, domain priors, and hand-crafted rules perform well in the short term, but in the long run are always surpassed by general methods that rely solely on compute and data.
50%UnverifiedDyna-Q was proposed by reinforcement learning pioneer Richard Sutton in 1991 and is a method that combines model-based learning with model-free learning.
50%Unverified《Reinforcement Learning: An Introduction》由Richard Sutton和Andrew Barto合著，第二版于2018年更新，全书可免费在线阅读
50%Unverified时序差分（TD）学习由 Richard Sutton 在1988年正式提出，结合了蒙特卡洛方法和动态规划的优点
50%