Multi-task reinforcement learning in humans
The ability to transfer knowledge across tasks and generalize to novel ones is an important hallmark of human intelligence. Yet not much is known about human multitask reinforcement learning. We study participants’ behaviour in a two-step decision-making task with multiple features and changing rewa...
Saved in:
Published in | Nature human behaviour Vol. 5; no. 6; pp. 764 - 773 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
London
Nature Publishing Group UK
01.06.2021
Nature Publishing Group |
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The ability to transfer knowledge across tasks and generalize to novel ones is an important hallmark of human intelligence. Yet not much is known about human multitask reinforcement learning. We study participants’ behaviour in a two-step decision-making task with multiple features and changing reward functions. We compare their behaviour with two algorithms for multitask reinforcement learning, one that maps previous policies and encountered features to new reward functions and one that approximates value functions across tasks, as well as to standard model-based and model-free algorithms. Across three exploratory experiments and a large preregistered confirmatory experiment, our results provide evidence that participants who are able to learn the task use a strategy that maps previously learned policies to novel scenarios. These results enrich our understanding of human reinforcement learning in complex environments with changing task demands.
Studying behaviour in a decision-making task with multiple features and changing reward functions, Tomov et al. find that a strategy that combines successor features with generalized policy iteration predicts behaviour best. |
---|---|
Bibliography: | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
ISSN: | 2397-3374 2397-3374 |
DOI: | 10.1038/s41562-020-01035-y |