Combining Denoised Neural Network and Genetic Symbolic Regression for Memory Behavior Modeling via Dynamic Asynchronous Optimization
Keywords: Memory behavior, asynchronous optimization, neural networks, genetic symbolic regression
TL;DR: The self-evolving psychological knowledge-informed neural network model (SPsyINN) uses Genetic Symbolic Regression to improve classical memory equations and enhances predictive accuracy and interpretability through neural networks.
Abstract: Memory behavior modeling is a critical topic in cognitive psychology and education. Traditional psychological approaches describe the dynamic properties of memory through memory equations derived from experimental data, but these models often lack accuracy and are frequently debated in terms of their form. In recent years, data-driven modeling methods have improved predictive accuracy but often suffer from poor interpretability, limiting their ability to provide deeper cognitive insights. While knowledge-informed neural network models have achieved significant success in fields such as physics, their application in behavior modeling remains limited. This paper proposes a Self-evolving Psychology-informed Neural Network (SPsyINN), which leverages classical memory equations as knowledge modules to constrain neural network training. To address challenges such as the difficulty in quantifying descriptors and the limited interpretability of classical memory equations, a genetic symbolic regression algorithm is introduced to conduct evolutionary searches for more optimal expressions based on classical memory equations, enabling the mutual progress of the knowledge module and the neural network module. Specifically, the proposed approach combines genetic symbolic regression and neural networks in a parallel training framework, with a dynamic joint optimization loss function ensuring effective knowledge alignment between the two modules. Then, for addressing the training efficiency differences arising from the distinct optimization methods and computational hardware requirements of genetic algorithms and neural networks, an asynchronous interaction mechanism mediated by proxy data is developed to facilitate effective communication between modules and improve optimization efficiency. Finally, a denoising module is integrated into the neural network to enhance robustness against data noise and improve generalization performance. Experimental results on four large-scale real-world memory behavior demonstrate that SPsyINN outperforms state-of-the-art methods in predictive accuracy. Ablation studies further show that the proposed approach effectively achieves mutual progress between different modules, improving model predictive accuracy while uncovering more interpretable memory equations, highlighting the potential application value of SPsyINN in psychological research. Our code is released at: \href{https://anonymous.4open.science/r/SPsyINN-3F18}{https://anonymous.4open.science/r/SPsyINN-3F18}
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.
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: 8572
Loading