DS$^2$-ABSA: Dual-Stream Data Synthesis with Label Refinement for Few-Shot Aspect-Based Sentiment Analysis

TL;DR

This paper introduces DS$^2$-ABSA, a dual-stream data synthesis framework utilizing large language models to generate diverse, high-quality data with refined labels for few-shot aspect-based sentiment analysis, significantly improving performance.

Contribution

The paper presents a novel dual-stream data synthesis approach with label refinement for few-shot ABSA, addressing diversity and label accuracy issues in existing LLM-based data augmentation methods.

Findings

Outperforms previous few-shot ABSA methods

Generates diverse and high-quality synthetic data

Achieves significant performance improvements

Abstract

Recently developed large language models (LLMs) have presented promising new avenues to address data scarcity in low-resource scenarios. In few-shot aspect-based sentiment analysis (ABSA), previous efforts have explored data augmentation techniques, which prompt LLMs to generate new samples by modifying existing ones. However, these methods fail to produce adequately diverse data, impairing their effectiveness. Besides, some studies apply in-context learning for ABSA by using specific instructions and a few selected examples as prompts. Though promising, LLMs often yield labels that deviate from task requirements. To overcome these limitations, we propose DS$^2$-ABSA, a dual-stream data synthesis framework targeted for few-shot ABSA. It leverages LLMs to synthesize data from two complementary perspectives: \textit{key-point-driven} and \textit{instance-driven}, which effectively generate diverse and high-quality ABSA samples in low-resource settings. Furthermore, a \textit{label refinement} module is integrated to improve the synthetic labels. Extensive experiments demonstrate that DS$^2$-ABSA significantly outperforms previous few-shot ABSA solutions and other LLM-oriented data generation methods.