Interactive Reinforcement Learning from Imperfect Teachers
| Autor: | Andrea L. Thomaz, Taylor Kessler Faulkner |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Computer science
media_common.quotation_subject 02 engineering and technology 010501 environmental sciences 01 natural sciences Robot learning Attention span Human–robot interaction Work (electrical) Human–computer interaction 0202 electrical engineering electronic engineering information engineering Reinforcement learning Robot 020201 artificial intelligence & image processing Quality (business) Imperfect 0105 earth and related environmental sciences media_common |
| Zdroj: | HRI (Companion) |
| DOI: | 10.1145/3434074.3446361 |
| Popis: | Robots can use information from people to improve learning speed or quality. However, people can have short attention spans and misunderstand tasks. Our work addresses these issues with algorithms for learning from inattentive teachers that take advantage of feedback when people are present, and an algorithm for learning from inaccurate teachers that estimates which state-action pairs receive incorrect feedback. These advances will enhance robots' ability to take advantage of imperfect feedback from human teachers. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|