Machine Reinforced Perturbation on Drifted Human Logical Reasoning
Keywords: Human Logical Reasoning, Deep Reinforcement Learning, Cognitive Model
TL;DR: This paper integrates data-driven deep reinforcement learning with drift-diffusion in cognitive science to simulate dynamic human logical reasoning response to environmental stimuli.
Abstract: Using deep neural networks as computational models to simulate cognitive process can provide key insights into human behavioral dynamics. This enables synthetic data generation to test hypotheses for neuroscience and guides adaptive interventions for cognitive regulation. Challenges arise when environments are highly dynamic, obscuring stimulus-behavior relationships. However, the majority of current research focuses on simulating human cognitive behaviors under ideal conditions, neglecting the influence of environmental disturbances. We propose ReactiveAgent, integrating drift-diffusion with deep reinforcement learning to simulate granular effects of dynamic environmental stimuli on human logical reasoning process. This framework is built and evaluated upon our contributed large dataset of 21,157 logical responses of humans under various dynamic stimuli. Quantitatively, the framework improves cognition modelling by considering temporal effect of environmental stimuli on logical reasoning and captures both subject-specific and stimuli-specific behavioural differences. Qualitatively, it captures general trends in human logical reasoning under stress, better than baselines. Our approach is extensible to examining diverse environmental influences on cognitive behaviors. Overall, it demonstrates a powerful, data-driven methodology to simulate, align with, and understand the vagaries of human logical reasoning in dynamic contexts.
Primary Area: applications to neuroscience & cognitive science
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 4842
Loading