Brent Harrison
Papers
2
Total Citations
86
H-Index
2
About
Brent Harrison is a leading researcher in artificial intelligence, specializing in reinforcement learning (RL) and human-robot interaction. His core work focuses on bridging the gap between non-expert human instruction and machine learning, enabling robots to learn from natural, intuitive communication rather than requiring rigid, technical vocabularies. Harrison’s most influential contribution is his 2016 paper, “Learning From Explanations Using Sentiment and Advice in RL,” which has garnered 83 citations. In this work, he pioneered techniques that allow RL agents to interpret human advice and sentiment—such as praise or criticism—without needing explicit state information, making machine learning accessible to everyday users. He further advanced this paradigm with “Object-Focused Advice in Reinforcement Learning,” where he demonstrated how agents can learn from simple, object-oriented sentences, a significant step toward more natural human-robot teaching. Harrison’s research is pivotal for creating intelligent agents that can be trained by anyone, not just experts, with broad implications for assistive robotics, autonomous systems, and interactive AI. His work stands out for its focus on making AI learning as intuitive as teaching another person.
Research Focus
Key Achievements
Top Papers
- 1Learning From Explanations Using Sentiment and Advice in RL83 citations · 2016
- 2Object-Focused Advice in Reinforcement Learning3 citations · 2016