Aspect-Based Sentiment Analysis (ABSA)
156 papers with code • 18 benchmarks • 18 datasets
Aspect-Based Sentiment Analysis (ABSA) is a Natural Language Processing task that aims to identify and extract the sentiment of specific aspects or components of a product or service. ABSA typically involves a multi-step process that begins with identifying the aspects or features of the product or service that are being discussed in the text. This is followed by sentiment analysis, where the sentiment polarity (positive, negative, or neutral) is assigned to each aspect based on the context of the sentence or document. Finally, the results are aggregated to provide an overall sentiment for each aspect.
And recent works propose more challenging ABSA tasks to predict sentiment triplets or quadruplets (Chen et al., 2022), the most influential of which are ASTE (Peng et al., 2020; Zhai et al., 2022), TASD (Wan et al., 2020), ASQP (Zhang et al., 2021a) and ACOS with an emphasis on the implicit aspects or opinions (Cai et al., 2020a).
Target-dependent sentiment classification remains a challenge: modeling the semantic relatedness of a target with its context words in a sentence.
Such importance degree and text representation are calculated with multiple computational layers, each of which is a neural attention model over an external memory.
Aspect-based sentiment analysis (ABSA), which aims to identify fine-grained opinion polarity towards a specific aspect, is a challenging subtask of sentiment analysis (SA).
In this paper, we introduce a new subtask under ABSA, named aspect sentiment triplet extraction (ASTE).
A Multi-task Learning Model for Chinese-oriented Aspect Polarity Classification and Aspect Term Extraction
Aspect-based sentiment analysis (ABSA) task is a multi-grained task of natural language processing and consists of two subtasks: aspect term extraction (ATE) and aspect polarity classification (APC).
In this paper, we argue that both targets and contexts deserve special treatment and need to be learned their own representations via interactive learning.
In this work, we apply adversarial training, which was put forward by Goodfellow et al. (2014), to the post-trained BERT (BERT-PT) language model proposed by Xu et al. (2019) on the two major tasks of Aspect Extraction and Aspect Sentiment Classification in sentiment analysis.
Unlike topic models which typically assume independently generated words, word embedding models encourage words that appear in similar contexts to be located close to each other in the embedding space.