Bayesian Distributional Models of Executive Functioning

2025-09-30

AI Summary

To address data sparsity and unstable parameter estimation in executive function (EF) assessment, this paper proposes the DLVM-DALE joint framework. It integrates a Deep Latent Variable Model (DLVM) for Bayesian joint inference across tasks and participants, enhancing robustness and convergence speed of EF distribution parameter estimation under small-sample conditions. Complementing DLVM, the framework incorporates Data-Adaptive Active Learning via Entropy (DALE), an information-gain-driven strategy that dynamically optimizes test item selection. Compared to conventional fixed-order paradigms and random sampling, DALE achieves significantly higher estimation accuracy within the first 80 trials. DLVM outperforms iterative maximum likelihood estimation (IMLE) in both accuracy and convergence rate with limited data. By transcending static testing constraints, DLVM-DALE establishes a scalable, Bayesian active measurement paradigm for efficient, individualized EF assessment.

Technology Category

• Machine Learning: Active Learning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Estimation (IMLE). DLVM integrates observations across multiple executive function tasks and individuals, allowing parameter estimation even under sparse or incomplete data conditions. DLVM consistently outperformed IMLE, especially under with smaller amounts of data, and converges faster to highly accurate estimates of the true distributions. In a second set of analyses, DALE adaptively guided sampling to maximize information gain, outperforming random sampling and fixed test batteries, particularly within the first 80 trials. These findings establish the advantages of combining DLVMs cross-task inference with DALEs optimal adaptive sampling, providing a principled basis for more efficient cognitive assessments.

Problem

• Modeling executive functioning across tasks and individuals
• Estimating parameters with sparse or incomplete data
• Optimizing adaptive sampling for efficient cognitive assessments

Innovation

• DLVM integrates multi-task observations for parameter estimation
• DLVM outperforms IMLE with sparse data and faster convergence
• DALE uses adaptive sampling to maximize information gain