Artificial neural networks for model identification and parameter estimation in computational cognitive models

Milena Rmus,Ti-Fen Pan,Liyu Xia,Anne G. E. Collins

DOI: https://doi.org/10.1101/2023.09.14.557793

2024-04-02

Abstract:Computational cognitive models have been used extensively to formalize cognitive processes. Model parameters offer a simple way to quantify individual differences in how humans process information. Similarly, model comparison allows researchers to identify which theories, embedded in different models, provide the best accounts of the data. Cognitive modeling uses statistical tools to quantitatively relate models to data that often rely on computing/estimating the likelihood of the data under the model. However, this likelihood is computationally intractable for a substantial number of models. These relevant models may embody reasonable theories of cognition, but are often under-explored due to the limited range of tools available to relate them to data. We contribute to filling this gap in a simple way using artificial neural networks (ANNs) to map data directly onto model identity and parameters, bypassing the likelihood estimation. We test our instantiation of an ANN as a cognitive model fitting tool on classes of cognitive models with strong inter-trial dependencies (such as reinforcement learning models), which offer unique challenges to most methods. We show that we can adequately perform both parameter estimation and model identification using our ANN approach, including for models that cannot be fit using traditional likelihood-based methods. We further discuss our work in the context of the ongoing research leveraging simulation-based approaches to parameter estimation and model identification, and how these approaches broaden the class of cognitive models researchers can quantitatively investigate.

Animal Behavior and Cognition

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address the challenges of parameter estimation and model identification in computational cognitive models. Specifically: 1. **Computational Complexity**: The likelihood functions of many computational cognitive models are difficult to compute or approximate, which limits researchers' application and exploration of these models. For example, when observed data (such as choices) depend on unobserved variables (such as latent rules or participants' attention states), computing the likelihood function becomes very complex. 2. **Model Applicability**: Existing statistical tools (such as maximum likelihood estimation, Bayesian methods, etc.) are mainly applicable to models with computable likelihood functions. However, many reasonable cognitive theoretical models cannot be widely studied and applied due to computational complexity. 3. **Model Comparison**: In model comparison, computing the likelihood function is an important step in evaluating the goodness of fit of a model. For models with uncomputable likelihood functions, traditional model comparison methods often cannot be effectively applied. To address these issues, the authors propose a method based on artificial neural networks (ANN) that can directly map data to model identity and parameters, thereby bypassing the computation or approximation of the likelihood function. This method is applicable not only to models with computable likelihood functions but also to those with uncomputable likelihood functions. Through this method, researchers can more broadly test and validate different cognitive theoretical models, enhancing the applicability and explanatory power of the models.

Artificial neural networks for model identification and parameter estimation in computational cognitive models

Predictive and Interpretable: Combining Artificial Neural Networks and Classic Cognitive Models to Understand Human Learning and Decision Making

Learning Cognitive Models using Neural Networks

Discovering Cognitive Strategies with Tiny Recurrent Neural Networks

On Model-guided Neural Networks for System Identification

System identification of neural systems: If we got it right, would we know?

Biological constraints on neural network models of cognitive function

Towards Neural Network based Cognitive Models of Dynamic Decision-Making by Humans

Deep Learning-Based Parameter Estimation for Neurophysiological Models of Neuroimaging Data

Robust parameter estimation and identifiability analysis with Hybrid Neural Ordinary Differential Equations in Computational Biology

BayCANN: Streamlining Bayesian Calibration with Artificial Neural Network Metamodeling

The neuroconnectionist research programme

Neural Networks, Artificial Intelligence and the Computational Brain

Identifying relationships between cognitive processes across tasks, contexts, and time

Online Model Checking Approach Based Parameter Estimation to A Neuronal Fate Decision Simulation Model in Caenorhabditis Elegans with Hybrid Functional Petri Net with Extension

Designing a set of criteria for evaluating artificial neural networks trained with physics-based data to replicate molecular dynamics and other particle method trajectories

Neural Networks for Parameter Estimation in Intractable Models

Understanding Human Cognition Through Computational Modeling

The Importance of Constraint Smoothness for Parameter Estimation in Computational Cognitive Modeling

How can deep learning advance computational modeling of sensory information processing?

A neural network paradigm for modeling psychometric data and estimating IRT model parameters: Cross estimation network