Hierarchical Learning for Maze Navigation: Emergence of Mental Representations via Second-Order Learning

2025-09-17

AI Summary

This work investigates how second-order learning drives the emergence of structured mental representations—i.e., internal cognitive models isomorphic to environmental structure—in maze navigation. Addressing poor generalization in high-level cognition, we propose a hierarchical architecture: a graph convolutional network (GCN) serves as the first-order learner for path prediction, while a multilayer perceptron (MLP) acts as the second-order learner that dynamically modulates GCN parameters, enabling adaptive alignment of learning mechanisms with environmental structure. Experiments demonstrate significant improvements in navigation performance and robustness on unseen mazes. Crucially, this study provides the first empirical evidence that second-order learning enhances generalization by regulating the representational generation process to foster environment–cognition structural isomorphism. Our findings reveal that plasticity in learning mechanisms plays a pivotal role in shaping structured mental representations.

Technology Category

• Machine Learning: Structured Learning
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Computer Vision: Representation Learning for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Mental representation, characterized by structured internal models mirroring external environments, is fundamental to advanced cognition but remains challenging to investigate empirically. Existing theory hypothesizes that second-order learning -- learning mechanisms that adapt first-order learning (i.e., learning about the task/domain) -- promotes the emergence of such environment-cognition isomorphism. In this paper, we empirically validate this hypothesis by proposing a hierarchical architecture comprising a Graph Convolutional Network (GCN) as a first-order learner and an MLP controller as a second-order learner. The GCN directly maps node-level features to predictions of optimal navigation paths, while the MLP dynamically adapts the GCN's parameters when confronting structurally novel maze environments. We demonstrate that second-order learning is particularly effective when the cognitive system develops an internal mental map structurally isomorphic to the environment. Quantitative and qualitative results highlight significant performance improvements and robust generalization on unseen maze tasks, providing empirical support for the pivotal role of structured mental representations in maximizing the effectiveness of second-order learning.

Problem

• Empirically validate second-order learning emergence mental representations
• Hierarchical architecture adapts first-order learner novel maze environments
• Demonstrate structural isomorphism improves navigation performance generalization

Innovation

• Hierarchical architecture with GCN and MLP
• Second-order learning adapts first-order parameters
• Mental map isomorphism enables robust generalization