Solving Aspect Category Sentiment Analysis as a Text Generation Task

TL;DR

This paper proposes a novel approach to aspect category sentiment analysis by framing it as a text generation task, leveraging pre-trained seq2seq models for improved performance especially in few-shot and zero-shot scenarios.

Contribution

It introduces a direct text generation formulation for ACSA, enabling better utilization of pre-trained language models compared to traditional classification-based methods.

Findings

Achieves state-of-the-art results on multiple benchmarks.

Shows significant improvements in few-shot and zero-shot settings.

Demonstrates the effectiveness of generative framing for ACSA.

Abstract

Aspect category sentiment analysis has attracted increasing research attention. The dominant methods make use of pre-trained language models by learning effective aspect category-specific representations, and adding specific output layers to its pre-trained representation. We consider a more direct way of making use of pre-trained language models, by casting the ACSA tasks into natural language generation tasks, using natural language sentences to represent the output. Our method allows more direct use of pre-trained knowledge in seq2seq language models by directly following the task setting during pre-training. Experiments on several benchmarks show that our method gives the best reported results, having large advantages in few-shot and zero-shot settings.