Aspect-Sentiment-Multiple-Opinion Triplet Extraction

Aspect Sentiment Triplet Extraction (ASTE) aims to extract aspect term (aspect), sentiment and opinion (opinion) triplets from sentences can tell a complete story, i.e., the discussed aspect, toward cause of sentiment. ASTE is charming task, however, one triplet extracted by only includes but an in sentence may have multiple corresponding opinions provides part reason why has this sentiment, as consequence, some are hard understand, provide erroneous information for downstream tasks. In paper, we introduce new named Multiple Opinions (ASMOTE). ASMOTE triplets. Specifically, contains all about exact that We propose Aspect-Guided Framework (AGF) address task. AGF first extracts aspects, then predicts their sentiments. Moreover, with help proposed Sequence Labeling Attention (SLA), improves performance classification using opinions. Experimental results on datasets demonstrate effectiveness our approach (Data code be found at https://github.com/l294265421/ASMOTE).