One Model, Two Minds: A Context-Gated Graph Learner that Recreates Human Biases

2025-09-10

AI Summary

This work addresses the challenge of developing human-like AI agents capable of emulating dual-process cognition—intuitive System 1 and deliberative System 2 reasoning—as well as characteristic cognitive biases (e.g., anchoring, framing, priming, and cognitive load-induced fatigue). Methodologically, we propose a cognitively grounded dual-process framework: graph convolutional networks model rapid, associative System 1 inference; meta-learning implements context-sensitive, reflective System 2 reasoning; and a learnable contextual gating mechanism dynamically orchestrates their interaction. To our knowledge, this is the first unified model to systematically reproduce multiple social-cognitive biases within a single architecture. It accurately fits human response patterns in canonical false-belief tasks, demonstrating strong generalization across novel contexts and high behavioral fidelity to human cognition. The framework provides an interpretable, controllable computational foundation for human-like social cognition in AI.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Humans and AI: Learning Human Values and Preferences
• Multiagent Systems: Adversarial Agents

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

We introduce a novel Theory of Mind (ToM) framework inspired by dual-process theories from cognitive science, integrating a fast, habitual graph-based reasoning system (System 1), implemented via graph convolutional networks (GCNs), and a slower, context-sensitive meta-adaptive learning system (System 2), driven by meta-learning techniques. Our model dynamically balances intuitive and deliberative reasoning through a learned context gate mechanism. We validate our architecture on canonical false-belief tasks and systematically explore its capacity to replicate hallmark cognitive biases associated with dual-process theory, including anchoring, cognitive-load fatigue, framing effects, and priming effects. Experimental results demonstrate that our dual-process approach closely mirrors human adaptive behavior, achieves robust generalization to unseen contexts, and elucidates cognitive mechanisms underlying reasoning biases. This work bridges artificial intelligence and cognitive theory, paving the way for AI systems exhibiting nuanced, human-like social cognition and adaptive decision-making capabilities.

Problem

• Modeling dual-process reasoning to replicate human cognitive biases
• Balancing intuitive and deliberative reasoning through context gates
• Validating on false-belief tasks to mirror adaptive human behavior

Innovation

• Dual-process graph learner with GCNs
• Meta-adaptive learning system with context gate
• Replicates human cognitive biases through balancing mechanism