English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
Add to Basket
 Local TagsRelease HistoryDetailsSummary
  Compositional generalization in multi-armed bandits

Saanum, T., Schulz, E., & Speekenbrink, M. (2021). Compositional generalization in multi-armed bandits. In 43rd Annual Conference of the Cognitive Science Society (CogSci 2021): Workshop 2 Using Games to Understand Intelligence (pp. 1320-1326). Red Hook, NY, USA: Curran.

Item is

 

show all hide all

Basic

show hide
Item Permalink: https://hdl.handle.net/21.11116/0000-0007-FF3C-8 Version Permalink: https://hdl.handle.net/21.11116/0000-000A-E9EF-3
Genre: Conference Paper

Files

show Files

Creators

show
hide
 Creators:
Saanum, T1, 2, Author           
Schulz, E1, 2, Author           
Speekenbrink, M, Author
Affiliations:
1Research Group Computational Principles of Intelligence, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_3189356              
2Max Planck Institute for Biological Cybernetics, Max Planck Society, Spemannstrasse 38, 72076 Tübingen, DE, ou_1497794              

Content

show
hide
Free keywords: -
 Abstract: To what extent do human reward learning and decision-making
rely on the ability to represent and generate richly structured
relationships between options? We provide evidence that
structure learning and the principle of compositionality play
crucial roles in human reinforcement learning. In a new multiarmed
bandit paradigm, we found evidence that participants
are able to learn representations of different reward structures
and combine them to make correct generalizations about options
in novel contexts. Moreover, we found substantial evidence
that participants transferred knowledge of simpler reward
structures to make compositional generalizations about
rewards in complex contexts. This allowed participants to
accumulate more rewards earlier, and to explore less whenever
such knowledge transfer was possible. We also provide
a computational model which is able to generalize and compose
knowledge for complex reward structures. This model describes
participant behaviour in the compositional generalization
task better than various other models of decision-making
and transfer learning.

Details

show
hide
Language(s):
 Dates: 2021-072021-11
 Publication Status: Issued
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: -
 Degree: -

Event

show
hide
Title: 43rd Annual Conference of the Cognitive Science Society (CogSci 2021)
Place of Event: Wien, Austria
Start-/End Date: 2021-07-26 - 2021-07-29

Legal Case

show

Project information

show

Source 1

show
hide
Title: 43rd Annual Conference of the Cognitive Science Society (CogSci 2021): Workshop 2 Using Games to Understand Intelligence
Source Genre: Proceedings
 Creator(s):
Affiliations:
Publ. Info: Red Hook, NY, USA : Curran
Pages: - Volume / Issue: - Sequence Number: - Start / End Page: 1320 - 1326 Identifier: ISBN: 978-1-7138-3525-7