Exploitative and Exploratory Attention in a Four-Armed Bandit Task

Walker, Adrian R. and Le Pelley, Mike E. and Beesley, Tom (2017) Exploitative and Exploratory Attention in a Four-Armed Bandit Task. In: CogSci 2017 - Proceedings of the 39th Annual Meeting of the Cognitive Science Society : Computational Foundations of Cognition. CogSci 2017 - Proceedings of the 39th Annual Meeting of the Cognitive Science Society: Computational Foundations of Cognition . The Cognitive Science Society, GBR, pp. 3484-3489. ISBN 9780991196760

Full text not available from this repository.

Abstract

When making decisions, we are often forced to choose between something safe we have chosen before, and something unknown to us that is inherently risky, but may provide a better long-term outcome. This problem is known as the Exploitation-Exploration (EE) Trade-Off. Most previous studies on the EE Trade-Off have relied on response data, leading to some ambiguity over whether uncertainty leads to true exploratory behavior, or whether the pattern of responding simply reflects a simpler ratio choice rule (such as the Generalized Matching Law (Baum, 1974; Herrnstein, 1961)). Here, we argue that the study of this issue can be enriched by measuring changes in attention (via eye-gaze), with the potential to disambiguate these two accounts. We find that when moving from certainty into uncertainty, the overall level of attention to stimuli in the task increases; a finding we argue is outside of the scope of ratio choice rules.

Item Type:
Contribution in Book/Report/Proceedings
Additional Information:
Publisher Copyright: © CogSci 2017.
Uncontrolled Keywords:
/dk/atira/pure/subjectarea/asjc/1700/1702
Subjects:
?? exploration trade-offreinforcement learningartificial intelligencecomputer science applicationshuman-computer interactioncognitive neuroscience ??
ID Code:
216975
Deposited By:
Deposited On:
30 Apr 2024 12:20
Refereed?:
Yes
Published?:
Published
Last Modified:
16 Jul 2024 05:27