Single Ground Truth Is Not Enough: Adding Flexibility to Aspect-Based Sentiment Analysis Evaluation
This work addresses the evaluation bottleneck in ABSA for researchers and practitioners by providing a more flexible and cost-effective framework, though it is incremental as it builds on existing evaluation methods.
The paper tackles the problem of evaluating aspect-based sentiment analysis (ABSA) by addressing the variability in span annotations due to subjectivity, proposing an automated pipeline to expand evaluation sets with alternative valid terms, which improves human agreement by up to 10 percentage points in Kendall's Tau score.
Aspect-based sentiment analysis (ABSA) is a challenging task of extracting sentiments along with their corresponding aspects and opinion terms from the text. The inherent subjectivity of span annotation makes variability in the surface forms of extracted terms, complicating the evaluation process. Traditional evaluation methods often constrain ground truths (GT) to a single term, potentially misrepresenting the accuracy of semantically valid predictions that differ in surface form. To address this limitation, we propose a novel and fully automated pipeline that expands existing evaluation sets by adding alternative valid terms for aspect and opinion. Our approach facilitates an equitable assessment of language models by accommodating multiple-answer candidates, resulting in enhanced human agreement compared to single-answer test sets (achieving up to a 10\%p improvement in Kendall's Tau score). Experimental results demonstrate that our expanded evaluation set helps uncover the capabilities of large language models (LLMs) in ABSA tasks, which is concealed by the single-answer GT sets. Consequently, our work contributes to the development of a flexible evaluation framework for ABSA by embracing diverse surface forms to span extraction tasks in a cost-effective and reproducible manner. Our code and dataset is open at https://github.com/dudrrm/zoom-in-n-out-absa.