Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Heterogeneous Debate Engine: Identity-Grounded Cognitive Architecture for Resilient LLM-Based Ethical Tutoring

Jakub Masłowski, Jarosław A. Chudziak

Abstract

Large Language Models (LLMs) are being increasingly used as autonomous agents in complex reasoning tasks, opening the niche for dialectical interactions. However, Multi-Agent systems implemented with systematically unconstrained systems systematically undergo semantic drift and logical deterioration and thus can hardly be used in providing ethical tutoring where a precise answer is required. Current simulation often tends to degenerate into dialectical stagnation, the agents degenerate into recursive concurrence or circular arguments. A critical challenge remains: how to enforce doctrinal fidelity without suppressing the generative flexibility required for dialectical reasoning? To address this niche, we contribute the Heterogeneous Debate Engine (HDE), a cognitive architecture that combines Identity-Grounded Retrieval-Augmented Generation (ID-RAG) for doctrinal fidelity and Heuristic Theory of Mind for strategic opponent modeling. Our evaluation shows that architectural heterogeneity is a crucial variable to stability: contrary doctrinal initializations (e.g., Deontology vs. Utilitarianism) have increased the Argument Complexity Scores of students by an order of magnitude, over baselines. These findings validate the effectiveness of ID-RAG and Heuristic ToM as architectural requirements in maintaining high-fidelity (adversarial) pedagogy.

Heterogeneous Debate Engine: Identity-Grounded Cognitive Architecture for Resilient LLM-Based Ethical Tutoring

Abstract

Large Language Models (LLMs) are being increasingly used as autonomous agents in complex reasoning tasks, opening the niche for dialectical interactions. However, Multi-Agent systems implemented with systematically unconstrained systems systematically undergo semantic drift and logical deterioration and thus can hardly be used in providing ethical tutoring where a precise answer is required. Current simulation often tends to degenerate into dialectical stagnation, the agents degenerate into recursive concurrence or circular arguments. A critical challenge remains: how to enforce doctrinal fidelity without suppressing the generative flexibility required for dialectical reasoning? To address this niche, we contribute the Heterogeneous Debate Engine (HDE), a cognitive architecture that combines Identity-Grounded Retrieval-Augmented Generation (ID-RAG) for doctrinal fidelity and Heuristic Theory of Mind for strategic opponent modeling. Our evaluation shows that architectural heterogeneity is a crucial variable to stability: contrary doctrinal initializations (e.g., Deontology vs. Utilitarianism) have increased the Argument Complexity Scores of students by an order of magnitude, over baselines. These findings validate the effectiveness of ID-RAG and Heuristic ToM as architectural requirements in maintaining high-fidelity (adversarial) pedagogy.

Paper Structure

This paper contains 21 sections, 3 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Approach, Methods, and Tools
  4. Research Methodology and Technology Stack
  5. Evaluation Framework
  6. System Architecture
  7. Architectural Topology and Debate Flow
  8. Cognitive Modules
  9. Orchestration and Moderation
  10. Experimental Setup
  11. Results and Evaluation
  12. Platform Capabilities and Debate Generation
  13. System Resilience Analysis
  14. Architectural Ablation Study
  15. Evaluation of Pedagogical Effectiveness
  16. ...and 6 more sections

Figures (3)

  • Figure 1: An illustrative example of the fragment of debate generated by the Heterogeneous Debate Engine. The figure demonstrates two autonomous agents (Kant and Mill) maintaining the doctrinal faithfulness to expose the student to philosophical conflict.
  • Figure 2: Control flow of the League of Moral Minds under the LangGraph orchestration. The flow starts with a coordinated deliberation and progresses to the loop of the iterative debate. The mechanisms behind it make use of cognitive modules in a loop.
  • Figure 3: Complementary Contributions of ID-RAG and ToM-Lite. Bars show Doctrinal Accuracy and Cross-Referencing. While ID-RAG maximizes identity stability, and ToM improves engagement, the Full System achieves high performance on both metrics.