Listwise Preference Optimization with Element-wise Confusions for Aspect Sentiment Quad Prediction

TL;DR

This paper introduces a novel listwise preference optimization framework for aspect sentiment quad prediction, enhancing structural validity and interpretability by explicitly modeling element relationships and confusions.

Contribution

It proposes a unified reasoning-based generation approach with element-wise confusable candidate training, improving accuracy and interpretability in aspect sentiment quadruple prediction.

Findings

Significant accuracy improvements on four benchmark datasets.

Enhanced explanation consistency and relational coherence.

Effective modeling of element-wise confusions through syntactic and semantic proximity.

Abstract

Aspect sentiment quad prediction (ASQP) is inherently challenging to predict a structured quadruple with four core sentiment elements, including aspect term (a), aspect category (c), opinion term (o), and sentiment polarity (s). Prior methods relying on marker-based prediction struggle with modeling the intricate relationships among elements and experience sharp performance declines when predicting higher-order elements (e.g., c and s) under standard supervised fine-tuning. To address these limitations, we employ reasoning-based generation to output both the quadruple and a natural language rationale under element prefixes within a unified template, encouraging explicit relational reasoning and interpretability. To further enhance element-wise alignment, we introduce a listwise preference optimization framework for improving structural validity and relational coherence. Specifically, we generate element-wise confusable candidates via syntactic and semantic proximity, then train the model with listwise objectives to prefer the gold candidates over closely competing alternatives. Extensive experiments on four benchmark datasets demonstrate that our framework effectively improves quadruple prediction accuracy and explanation consistency.