Title
Reinforcement Learning-Based Persuasion by a Conversational Agent for Behavior Change
Author
Albers, N.
Neerincx, M.A.
Brinkman, W.P.
Contributor
Leiva, L.A. (editor)
Pruski, C. (editor)
Markovich, R. (editor)
Najjar, A. (editor)
Schommer, C. (editor)
Publication year
2021
Abstract
There is certainly some behavior that you want to change. Maybe you want to become more physically active, call your mother more often or snack less when watching TV at night. Let’s assume that you want to quit smoking. You are not doing this alone, but are supported by your coach Hannah. Hannah constantly persuades you to stick to your intervention. How does she decide how to do that? First, Hannah has a lot of theoretical expertise. Moreover, you are not Hannah’s first client, so she can draw upon her experience with other and especially similar clients. Third, Hannah considers your current situation - are you confident or stressed about a deadline? In addition, she will persuade you in such a way that she can persuade you again in the future. And lastly, Hannah will keep adapting her strategy over time. Now, let’s suppose that you have another coach, Sam. Unlike Hannah, Sam is a virtual coach. Can Sam do what Hannah can?
Subject
Conversational agent
Persuasion
Reinforcement learning
To reference this document use:
http://resolver.tudelft.nl/uuid:5ca8f3b1-8efc-4218-a891-0b6b3a94a344
TNO identifier
966573
Source
Proceedings of BNAIC/BeneLearn 2021 : 33rd Benelux Conference on Artificial Intelligence and 30th Belgian-Dutch Conference on Machine Learning : AI in Action, 729-732
Bibliographical note
Poster session
Document type
conference paper