Towards Robust Argumentative Essay Understanding via TIDE: An Interactive Framework with Trial and Debate

2026-05-17

AI Summary

This work addresses the susceptibility of existing methods to noisy data and the instability of prompt optimization in argumentative text understanding tasks. To this end, the authors propose TIDE, a novel framework that, for the first time, integrates trial-and-error and debate mechanisms into rule-based prompt optimization. By iteratively refining prompts and enabling multi-perspective interactions, TIDE enhances the model’s robust comprehension of argumentative structures. The approach effectively mitigates the adverse effects of noise and improves optimization stability. Experimental results demonstrate that TIDE significantly outperforms current state-of-the-art methods across three distinct tasks: automated essay scoring, argument component detection, and argument relation identification, thereby validating its effectiveness and generalizability.

Abstract

Argumentative essays serve as a vital medium for assessing critical thinking and reasoning skills, yet there is limited works on accurately understanding and evaluating such texts via prompt. In this work, we propose TIDE, a novel framework designed to improve criteria-based prompt optimization for argument-related tasks by integrating TrIal and DEbate mechanism. Our method addresses key limitations of criteria-based prompt optimizing by mitigating the influence of noisy training data and enhancing optimization stability. We evaluate TIDE on three core tasks: Automated Essay Scoring, Argument Component Detection, and Argument Relation Identification. Results demonstrate that our framework improves performance across tasks. These findings underscore the potential of combining prompt-based methods for advanced argument understanding.

Problem

• Argumentative Essay Understanding
• Prompt Optimization
• Noisy Training Data
• Optimization Stability
• Argument Evaluation

Innovation

• TIDE
• prompt optimization
• argument understanding
• trial and debate
• automated essay scoring