Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence

By Chi Sun, Luyao Huang, Xipeng Qiu. 2019

727 citations recorded by the source corpus.

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 construct an auxiliary sentence from the aspect and convert ABSA to a sentence-pair classification task, such as question answering (QA) and natural language inference (NLI). We fine-tune the pre-trained model from BERT and achieve new state-of-the-art results on SentiHood and SemEval-2014 Task 4 datasets. The source codes are available at https://github.com/HSLCY/ABSA-BERT-pair.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Bing Liu
2. Erik Cambria
3. Soujanya Poria
4. Zhiyuan Liu
5. Maosong Sun
6. Zixuan Ke
7. Rada Mihalcea
8. Hu Xu

In Natural Language Processing

1. Maosong Sun
2. Zhiyuan Liu
3. Soujanya Poria
4. Rada Mihalcea
5. Xipeng Qiu
6. Xuanjing Huang
7. Bing Liu
8. Hu Xu

Citations in the corpus

• Big Bird: Transformers for Longer Sequences (2020, methodology): “…based tasks – as a seq2seq model for translation [92], summarization [66], generation [15], etc. or as a standalone encoders for sentiment analysis [84], POS tagging [65], machine reading comprehension [94], etc. – and it is known to vastly outperform previous sequence models like LSTM [37].”
• Achieving Forgetting Prevention and Knowledge Transfer in Continual Learning (2021, methodology): “In NLP, fine-tuning a BERT [8] like pre-trained language model has been regarded as one of the most effective techniques in applications [65, 57].”
• Pre-trained models for natural language processing: A survey (2020, methodology): “Despite their success in simple sentiment classification, directly applying BERT to aspect-based sentiment analysis (ABSA), which is a fine-grained SA task, shows less significant improvement [166].”
• CLASSIC: Continual and Contrastive Learning of Aspect Sentiment Classification Tasks (2021, background): “Previous research has shown that one of the most effective approaches for ASC (Xu et al., 2019; Sun et al., 2019) is to fine-tune the BERT (Devlin et al.”
• Template-Based Named Entity Recognition Using BART (2021, methodology): “Sun et al. (2019) use templates to construct auxiliary sentences, and transform aspect sentiment task as a sentence-pair classification task.”
• Beneath the Tip of the Iceberg: Current Challenges and New Directions in Sentiment Analysis Research (2020, background): “What remains to be seen is the future role of BERT-based networks working in conjunction with the task-dependent designs existing as the present state of the arts in this area (Sun et al., 2019).”
• Sentiment Analysis for E-Commerce Product Reviews in Chinese Based on Sentiment Lexicon and Deep Learning (2020, methodology): “Commonly used word vector implementation models are Word2Vec [51], Glove [52], ELMo [53] and BERT [54].”
• Target-Dependent Sentiment Classification With BERT (2019, methodology): “Unfortunately, [36] focuses on the solution of subtask 4”
• Aspect-Based Sentiment Analysis using BERT (2019, result): “Solving ABSA as a sentence-pair classification task using BERT by constructing auxiliary sentence has been seen to improve the results, compared to the previous state-of-the-art models that used single-sentence classification (Sun et al., 2019).”
• IndoLEM and IndoBERT: A Benchmark Dataset and Pre-trained Language Model for Indonesian NLP (2020, background): “, 2019) and sentence semantics of text (Sun et al., 2019).”
• Aspect Sentiment Quad Prediction as Paraphrase Generation (2021, neutral): “Sun et al. (2019) transform the aspect sentiment classification task as a language inference problem by constructing an auxiliary sentence.”
• Knowledge-Grounded Dialogue Generation with Pre-trained Language Models (2020, background): “…are revolutionizing almost the full spectrum of NLP applications (Raffel et al., 2019; Sun et al., 2019b; Qiao et al., 2019; Zhang et al., 2019b; Lam-ple and Conneau, 2019) and some interdisciplinary applications in NLP and computer vision (Lu et al., 2019; Su et al., 2019; Sun et al., 2019a).”
• Syntax-Aware Aspect Level Sentiment Classification with Graph Attention Networks (2019, background): “Benefited from the rich linguistic knowledge learned from massive language modeling (Devlin et al., 2018), researchers show great progress on this task using BERT representations (Sun et al., 2019).”
• Natural Language Processing Advancements By Deep Learning: A Survey (2020, methodology): “tasks [69], post-training approaches [127] on the popular”
• Dual Graph Convolutional Networks for Aspect-based Sentiment Analysis (2021, methodology): “Following this line, a few works extend the GCN and GAT models by means of a syntactical dependency tree and develop several outstanding models (Zhang et al., 2019; Sun et al., 2019b; Huang and Carley, 2019; Wang et al., 2020; Tang et al., 2020).”
• Coreferential Reasoning Learning for Language Representation (2020, background): “…tasks, such as natural language inference, sentiment classification, question answering, relation extraction, fact extraction and verification, and coreference resolution (Zhang et al., 2019; Sun et al., 2019; Tal-mor and Berant, 2019; Peters et al., 2019; Zhou et al., 2019; Joshi et al., 2019b).”