Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

From Pre-trained Models to Large Language Models: A Comprehensive Survey of AI-Driven Psychological Computing

Huiyao Chen, Ruimeng Liu, Yan Luo, Jiawen Zhang, Meishan Zhang, Baotian Hu, Min Zhang

Abstract

The intersection of artificial intelligence and psychological science has experienced remarkable growth, with annual publications expanding from 859 papers in 2000 to 29,979 by 2025. However, this rapid evolution has created methodological fragmentation where similar computational techniques are independently developed across isolated psychological domains. This survey introduces the first systematic taxonomy that organizes AI-driven psychology tasks by computational processing patterns rather than application domains, categorizing them into four fundamental types: classification, regression, structured relational, and generative interactive tasks. Through analysis of over 300 representative works spanning the pre-trained model era and large language model era, we examine how computational approaches evolved from task-specific feature engineering to transfer learning and few-shot adaptation. We provide systematic coverage of datasets, evaluation metrics, and benchmarks while addressing fundamental challenges including interpretability, label uncertainty, privacy constraints, and cross-cultural validity. This computational perspective reveals transferable methodological patterns previously obscured by domain-centric organization, enabling systematic knowledge transfer and accelerated progress in computational psychology.

From Pre-trained Models to Large Language Models: A Comprehensive Survey of AI-Driven Psychological Computing

Abstract

The intersection of artificial intelligence and psychological science has experienced remarkable growth, with annual publications expanding from 859 papers in 2000 to 29,979 by 2025. However, this rapid evolution has created methodological fragmentation where similar computational techniques are independently developed across isolated psychological domains. This survey introduces the first systematic taxonomy that organizes AI-driven psychology tasks by computational processing patterns rather than application domains, categorizing them into four fundamental types: classification, regression, structured relational, and generative interactive tasks. Through analysis of over 300 representative works spanning the pre-trained model era and large language model era, we examine how computational approaches evolved from task-specific feature engineering to transfer learning and few-shot adaptation. We provide systematic coverage of datasets, evaluation metrics, and benchmarks while addressing fundamental challenges including interpretability, label uncertainty, privacy constraints, and cross-cultural validity. This computational perspective reveals transferable methodological patterns previously obscured by domain-centric organization, enabling systematic knowledge transfer and accelerated progress in computational psychology.

Paper Structure

This paper contains 103 sections, 33 equations, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. A Computational Perspective on Psychological Tasks
  3. Survey Scope and Contributions
  4. Background
  5. Characteristics of Psychological Computing
  6. Psychological Foundations
  7. Data Characteristics
  8. Technical Distinctions
  9. Formal Task Formulation
  10. A Computational Taxonomy
  11. Taxonomy Design Principles
  12. Classification Tasks
  13. Mental State Recognition
  14. Behavioral Pattern Classification
  15. Regression Tasks
  16. ...and 88 more sections

Figures (6)

  • Figure 1: Publication growth trajectory in AI-driven psychology research from 2000 to 2025, showing three distinct phases of evolution from foundational growth to exponential expansion.
  • Figure 2: Disciplinary shift in AI-driven psychology research across three developmental phases (2000-2025). The exponential growth in computer science publications (from $\sim$12K to $\sim$100K) compared to modest increases in traditional psychology domains illustrates the field's transformation from psychology-centric to computation-centric approaches.
  • Figure 3: Evolution of psychology methods. Conventional approaches rely on individual manual assessment by trained professionals, providing focused clinical insights through in-person consultations. AI-driven methods complement these practices by enabling automated, high-throughput assessment, precise quantification of psychological states, multi-source data integration, and accessible, anytime-anywhere support, thereby enhancing the efficiency, granularity, comprehensiveness, and availability of psychological services.
  • Figure 4: AI-driven psychology task computational framework. The figure illustrates four main task categories with their corresponding input modalities, network architectures, and output types: (a) Classification tasks for discrete psychological state identification, (b) Regression tasks for continuous psychological measurement, (c) Structured relational tasks for extracting psychological entities and relations as structured tuples, and (d) Generative interactive tasks for dynamic content creation and user engagement.
  • Figure 5: Comprehensive taxonomy of AI-driven psychological computing tasks.
  • ...and 1 more figures