Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Self-Consistent Reasoning-based Aspect-Sentiment Quad Prediction with Extract-Then-Assign Strategy

Jieyong Kim, Ryang Heo, Yongsik Seo, SeongKu Kang, Jinyoung Yeo, Dongha Lee

TL;DR

Self-Consistent Reasoning-based Aspect-sentiment quadruple Prediction (SCRAP) is proposed, optimizing its model to generate reasonings and the corresponding sentiment quadruplets in sequence, resulting in enhanced interpretability and accuracy in ASQP.

Abstract

In the task of aspect sentiment quad prediction (ASQP), generative methods for predicting sentiment quads have shown promising results. However, they still suffer from imprecise predictions and limited interpretability, caused by data scarcity and inadequate modeling of the quadruplet composition process. In this paper, we propose Self-Consistent Reasoning-based Aspect-sentiment quadruple Prediction (SCRAP), optimizing its model to generate reasonings and the corresponding sentiment quadruplets in sequence. SCRAP adopts the Extract-Then-Assign reasoning strategy, which closely mimics human cognition. In the end, SCRAP significantly improves the model's ability to handle complex reasoning tasks and correctly predict quadruplets through consistency voting, resulting in enhanced interpretability and accuracy in ASQP.

Self-Consistent Reasoning-based Aspect-Sentiment Quad Prediction with Extract-Then-Assign Strategy

TL;DR

Self-Consistent Reasoning-based Aspect-sentiment quadruple Prediction (SCRAP) is proposed, optimizing its model to generate reasonings and the corresponding sentiment quadruplets in sequence, resulting in enhanced interpretability and accuracy in ASQP.

Abstract

In the task of aspect sentiment quad prediction (ASQP), generative methods for predicting sentiment quads have shown promising results. However, they still suffer from imprecise predictions and limited interpretability, caused by data scarcity and inadequate modeling of the quadruplet composition process. In this paper, we propose Self-Consistent Reasoning-based Aspect-sentiment quadruple Prediction (SCRAP), optimizing its model to generate reasonings and the corresponding sentiment quadruplets in sequence. SCRAP adopts the Extract-Then-Assign reasoning strategy, which closely mimics human cognition. In the end, SCRAP significantly improves the model's ability to handle complex reasoning tasks and correctly predict quadruplets through consistency voting, resulting in enhanced interpretability and accuracy in ASQP.
Paper Structure (30 sections, 6 figures, 3 tables)

This paper contains 30 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. SCRAP Framework
  3. Problem Formulation
  4. ASQP Reasoning Generation
  5. Extract-Then-Assign reasoning
  6. Reasoning generation with LLM
  7. Supervised Fine-Tuning
  8. Target construction
  9. Training
  10. Self-Consistent Quad Prediction
  11. Experiments
  12. Experimental Settings
  13. Dataset and evaluation metrics
  14. Baselines
  15. Results and Discussion
  16. ...and 15 more sections

Figures (6)

  • Figure 1: An illustrative example of Extract-Then-Assign reasoning process for ASQP task.
  • Figure 1: ASQP performance comparison. Backbone model: T5-Base. The best and second-best results are in bold and underlined, respectively. $\dagger$ indicates the results reported from their original papers.
  • Figure 2: An overview of SCRAP which concurrently generates sentiment quads and the corresponding reasoning.
  • Figure 3: ASQP performance with T5-Base and T5-3B. Dataset: Rest15.
  • Figure 4: Error analysis and case study. Left: Analysis of prediction errors on the Rest16. We report the error rate for each element type of aspect sentiment quad. Middle and Right: The case study of SCRAP. We present the input sentence, gold quads, and the prediction made by SCRAP.
  • ...and 1 more figures