Learning to Represent Individual Differences for Choice Decision Making

2025-03-27

AI Summary

This study addresses the limited representational capacity of traditional individual-difference modeling in economic decision prediction—where reliance on low-dimensional questionnaires and rigid theoretical frameworks (e.g., prospect theory) hinders accuracy and interpretability. We propose the first deep representation learning framework tailored to individual differences, jointly encoding structured demographic features and unstructured free-text responses via a BERT variant. The model performs end-to-end optimization of decision-prediction objectives, yielding privacy-preserving, task-adaptive, open individual embeddings. Evaluated on a multi-task economic choice dataset, our approach achieves a 12.6% average accuracy gain over both baseline models without individual modeling and classical theoretical models, while demonstrating markedly improved generalization. This work establishes a novel paradigm for behavioral modeling: high-dimensional, scalable, and interpretable individual representations grounded in data-driven representation learning.

Technology Category

• Machine Learning: Representation Learning
• Game Theory and Economic Paradigms: Adversarial Learning
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Human decision making can be challenging to predict because decisions are affected by a number of complex factors. Adding to this complexity, decision-making processes can differ considerably between individuals, and methods aimed at predicting human decisions need to take individual differences into account. Behavioral science offers methods by which to measure individual differences (e.g., questionnaires, behavioral models), but these are often narrowed down to low dimensions and not tailored to specific prediction tasks. This paper investigates the use of representation learning to measure individual differences from behavioral experiment data. Representation learning offers a flexible approach to create individual embeddings from data that are both structured (e.g., demographic information) and unstructured (e.g., free text), where the flexibility provides more options for individual difference measures for personalization, e.g., free text responses may allow for open-ended questions that are less privacy-sensitive. In the current paper we use representation learning to characterize individual differences in human performance on an economic decision-making task. We demonstrate that models using representation learning to capture individual differences consistently improve decision predictions over models without representation learning, and even outperform well-known theory-based behavioral models used in these environments. Our results propose that representation learning offers a useful and flexible tool to capture individual differences.

Problem

• Predicting human decisions considering individual differences
• Measuring individual differences using representation learning
• Improving decision predictions with learned individual embeddings

Innovation

• Uses representation learning for individual differences
• Combines structured and unstructured data sources
• Outperforms traditional theory-based behavioral models