Apparently irrational choice as optimal sequential decision making

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Colleges, School and Institutes


In this paper, we propose a normative approach to modeling apparent human irrational decision making (cognitive biases) that makes use of inherently rational computational mechanisms. We view preferential choice tasks as sequential decision making problems and formulate them as Partially Observable Markov Decision Processes (POMDPs). The resulting sequential decision model learns what information to gather about which options, whether to calculate option values or make comparisons between options and when to make a choice. We apply the model to choice problems where con- text is known to influence human choice, an effect that has been taken as evidence that human cognition is irrational. Our results show that the new model approximates a bounded optimal cognitive policy and makes quantitative predictions that correspond well to evidence about human choice. Furthermore, the model learns to use context to help infer which option has a maximum expected value while taking into ac- count computational cost and cognitive limits. In addition, it predicts when, and explains why, people stop evidence accumulation and make a decision. We argue that the model pro- vides evidence that apparent human irrationalities are emergent consequences of learning processes that prefer higher value (rational) policies.


Original languageEnglish
Title of host publicationAAAI'21 Proceedings of the Thirty-fifth AAAI Conference on Artificial Intelligence
Publication statusPublished - 2 Feb 2021