Probabilities are central to rational accounts of human cognition. According to the Dutch book theorem, people are rational if and only if they obey the axioms of probability theory. From perception to memory and decision making, probabilistic approaches have inspired some of the most influential theories in cognitive science and contributed to the only two Nobel Prizes (Simon and Kahneman) awarded in the field.
Yet, decades of empirical work show that human probabilistic reasoning often violates basic axioms of probability theory (e.g., Tversky & Kahneman, 1983; Tversky & Koehler, 1994). Besides, research on fast and frugal heuristics has revealed that people frequently rely on simple deterministic rules rather than complex probabilistic models (Gigerenzer & Todd, 2000; Gigerenzer & Gaissmaier, 2011). These findings highlight a major research gap: the absence of a unified account that integrates probabilistic fallacies, rule-based reasoning, and probabilistic processes in human cognition.
My research addresses this gap through three main themes:
1. Understanding probability judgment errors and their role in decision making.
2. A unified account of probabilistic and heuristics reasoning.
3. Quantum probability as a new tool for cognitive modeling and artificial intelligence.
For the first theme, my work on the Quantum Sequential Sampler (Huang, Busemeyer, Ebelt & Pothos, 2024, Psychological Review) integrates axiomatic modifications to probability theory with sequential sampling models to explain systematic judgment errors. The model achieves state-of-the-art performance on the largest dataset of probability judgment errors to date and captures how such errors influence choice and response time dynamics. For the second theme, my ongoing dissertation develops a probabilistic extension of the fast and frugal heuristics framework, linking deterministic rules, rational linear regression, exemplar models, and stochastic decision processes. This extension allows probability judgment and response time predictions while jointly modeling various important behavioral effects.
The third theme builds on quantum cognition, a field pioneered by my graduate advisor Jerome Busemeyer, which applies quantum probability theory to model cognitive phenomena (Busemeyer & Bruza, 2012; Huang, Epping, Trueblood, Yearsley, Busemeyer & Pothos, 2025, Psychonomic Bulletin & Review). This framework shows promise in bridging the gap by providing unified accounts of behavioral effects such as the conjunction fallacy and order effects. Much of my past work has pursued this direction (e.g., Huang, Zhang, Xie, Breithaupt & Busemeyer, 2025, Cognition). Looking ahead, however, my ambition extends beyond demonstrating the “quantum-like” nature of cognition. I propose a new direction: leveraging this “quantum-like” character to develop “human-like quantum artificial intelligence.” This approach is agnostic to whether the human mind or brain is literally governed by quantum probability theory. Instead, it explores how quantum computing technology, combined with the insight that human bounded rationality often behaves in a “quantum-like" manner, can reduce computational complexity and implement bounded rational artificial intelligence algorithms (Pothukuchi, …, Huang et al., in press).
Together, these themes have three important broader impacts to cognitive science and artificial intelligence research. First, they unify accounts of bounded rationality, providing a more general framework for understanding human cognition. Second, they sharpen the meaning of bounded rationality in ways that inform the design of future human-like artificial intelligence. Finally, they deepen our understanding of how the “quantum-like” nature of cognition can be leveraged in artificial intelligence research, offering new opportunities which combine quantum computing and artificial intelligence.