ASCM4ABSA
ASCM4ABSA - Aspect-specific Context Modeling for Aspect-based Sentiment Analysis
- Our code and proposed data (adversarial ABSA benchmark, dubbed advABSA) for NLPCC 2022 paper titled "Aspect-specific Context Modeling for Aspect-based Sentiment Analysis"
- Fang Ma, Chen Zhang, Bo Zhang, and Dawei Song.
stdABSA (Standard ABSA Benchmark)
- stdABSA contains datasets from SemEval for aspect-based Sentiment Classification (SC) and Opinion Extraction (OE) evaluation, namely Sem-SC and Sem-OE.
advABSA (Adversarial ABSA Benchmark)
- Since there are only adversarial datasets for robust SC and not any for robust OE, we craft advABSA based on xing2020tasty's datasets and methods. That is, advABSA can be decomposed to two parts, where the first part is Arts-SC reused from xing2020tasty and the second part is Arts-OE produced by us.