Abstract
Aspect Sentiment Triplet Extraction (ASTE) is widely used in variousapplications. However, existing ASTE datasets are limited in their ability torepresent real-world scenarios, hindering the advancement of research in thisarea. In this paper, we introduce a new dataset, named DMASTE, which ismanually annotated to better fit real-world scenarios by providing more diverseand realistic reviews for the task. The dataset includes various lengths,diverse expressions, more aspect types, and more domains than existingdatasets. We conduct extensive experiments on DMASTE in multiple settings toevaluate previous ASTE approaches. Empirical results demonstrate that DMASTE isa more challenging ASTE dataset. Further analyses of in-domain and cross-domainsettings provide promising directions for future research. Our code and datasetare available at https://github.com/NJUNLP/DMASTE.