Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dialectical Reconciliation via Structured Argumentative Dialogues

Stylianos Loukas Vasileiou, Ashwin Kumar, William Yeoh, Tran Cao Son, Francesca Toni

TL;DR

DR-Arg introduces a dialectical reconciliation framework for explanation in human-AI interaction, leveraging structured deductive argumentation to resolve knowledge gaps without assuming a fixed human model. It replaces single-shot explanations with iterative dialogues where an explainer and explainee exchange arguments to achieve understanding of the explainer’s decisions, even if disagreements remain. The work provides formal operational semantics, termination and success guarantees, and an approximated explainee-understanding metric, complemented by computational simulations and a human-user study showing improved comprehension and satisfaction over single-shot approaches. Empirical results demonstrate that DR-Arg scales with knowledge-base size, yields greater understanding through iterative updates, and can outperform single-shot reconciliation in real-world-like tasks. Overall, the framework highlights the potential of argumentation-based dialogue to enhance explainability in critical domains where human-AI collaboration is essential.

Abstract

We present a novel framework designed to extend model reconciliation approaches, commonly used in human-aware planning, for enhanced human-AI interaction. By adopting a structured argumentation-based dialogue paradigm, our framework enables dialectical reconciliation to address knowledge discrepancies between an explainer (AI agent) and an explainee (human user), where the goal is for the explainee to understand the explainer's decision. We formally describe the operational semantics of our proposed framework, providing theoretical guarantees. We then evaluate the framework's efficacy ``in the wild'' via computational and human-subject experiments. Our findings suggest that our framework offers a promising direction for fostering effective human-AI interactions in domains where explainability is important.

Dialectical Reconciliation via Structured Argumentative Dialogues

TL;DR

DR-Arg introduces a dialectical reconciliation framework for explanation in human-AI interaction, leveraging structured deductive argumentation to resolve knowledge gaps without assuming a fixed human model. It replaces single-shot explanations with iterative dialogues where an explainer and explainee exchange arguments to achieve understanding of the explainer’s decisions, even if disagreements remain. The work provides formal operational semantics, termination and success guarantees, and an approximated explainee-understanding metric, complemented by computational simulations and a human-user study showing improved comprehension and satisfaction over single-shot approaches. Empirical results demonstrate that DR-Arg scales with knowledge-base size, yields greater understanding through iterative updates, and can outperform single-shot reconciliation in real-world-like tasks. Overall, the framework highlights the potential of argumentation-based dialogue to enhance explainability in critical domains where human-AI collaboration is essential.

Abstract

We present a novel framework designed to extend model reconciliation approaches, commonly used in human-aware planning, for enhanced human-AI interaction. By adopting a structured argumentation-based dialogue paradigm, our framework enables dialectical reconciliation to address knowledge discrepancies between an explainer (AI agent) and an explainee (human user), where the goal is for the explainee to understand the explainer's decision. We formally describe the operational semantics of our proposed framework, providing theoretical guarantees. We then evaluate the framework's efficacy ``in the wild'' via computational and human-subject experiments. Our findings suggest that our framework offers a promising direction for fostering effective human-AI interactions in domains where explainability is important.
Paper Structure (22 sections, 4 equations, 10 figures, 6 tables)

This paper contains 22 sections, 4 equations, 10 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Motivating Example
  3. Related Work
  4. Background: Deductive Argumentation
  5. DR-Arg Framework
  6. Key Assumptions:
  7. Dialectical Reconciliation Dialogue Type
  8. Agent Strategy:
  9. Operational Semantics of DR Dialogues
  10. Illustrative Example
  11. Properties of DR Dialogues
  12. Termination:
  13. Success:
  14. Approximating Explainee Understanding
  15. Empirical Evaluations
  16. ...and 7 more sections

Figures (10)

  • Figure 1: The introductory information shown to participants at the beginning of the study.
  • Figure 2: (DR-Arg) Initial interaction: starting query, response (support) to query, and follow-up questions (refute).
  • Figure 3: (DR-Arg) Second interaction: Refutation to user response, and follow-up questions.
  • Figure 4: (DR-Arg) Third interaction: Refutation to user response, and follow-up questions.
  • Figure 5: (DR-Arg) Fourth interaction: Refutation to user response, and follow-up questions.
  • ...and 5 more figures