ROAST: Review-level Opinion Aspect Sentiment Target Joint Detection for ABSA
Siva Uday Sampreeth Chebolu, Franck Dernoncourt, Nedim Lipka, Thamar Solorio
TL;DR
ROAST introduces a review-level ABSA task that jointly detects aspect categories, targets, opinions, and sentiment across entire reviews, bridging the gap between sentence- and text-level analyses. It extends ABSA datasets to multiple domains and languages with six new corpora, and evaluates a spectrum of baselines including MRC-based, BERT-based, and generative models, highlighting the strength of generation-based approaches like GAS and Paraphrase. The paper provides a rigorous annotation workflow, inter-annotator agreement metrics, and detailed statistics showing increased explicit targets/opinions when modeling full reviews. This work lays the groundwork for cross-domain and cross-lingual ABSA, enabling more realistic and scalable sentiment analysis in diverse languages and domains.
Abstract
Aspect-Based Sentiment Analysis (ABSA) has experienced tremendous expansion and diversity due to various shared tasks spanning several languages and fields and organized via SemEval workshops and Germeval. Nonetheless, a few shortcomings still need to be addressed, such as the lack of low-resource language evaluations and the emphasis on sentence-level analysis. To thoroughly assess ABSA techniques in the context of complete reviews, this research presents a novel task, Review-Level Opinion Aspect Sentiment Target (ROAST). ROAST seeks to close the gap between sentence-level and text-level ABSA by identifying every ABSA constituent at the review level. We extend the available datasets to enable ROAST, addressing the drawbacks noted in previous research by incorporating low-resource languages, numerous languages, and a variety of topics. Through this effort, ABSA research will be able to cover more ground and get a deeper comprehension of the task and its practical application in a variety of languages and domains (https://github.com/RiTUAL-UH/ROAST-ABSA).