References

Introduction to the BERT interpretability: 

• Analysis Methods in Neural Language Processing: A Survey 1812.08951
• Are Sixteen Heads Really Better than One? 1905.10650
• A Primer in BERTology: What we know about how BERT works 2002.12327
• What Does BERT Look At? An Analysis of BERT’s Attention 1906.04341
• Visualizing and Measuring the Geometry of BERT 1906.02715
• Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailment 1907.11932
• Adversarial Training for Aspect-Based Sentiment Analysis with BERT 2001.11316
• Adv-BERT: BERT is not robust on misspellings! Generating nature adversarial samples on BERT 2003.04985
• exBERT: A Visual Analysis Tool to Explore Learned Representations in Transformers Models 1910.05276
• Does BERT Make Any Sense? Interpretable Word Sense Disambiguation with Contextualized Embeddings 1909.10430)
• Attention is not Explanation 1902.10186
• Attention is not not Explanation 1908.04626
• Hierarchical interpretations for neural network predictions 1806.05337
• Analysis Methods in Neural NLP 10.1162

How to use language models in the Aspect-Based Sentiment Analysis: 

• Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence (NAACL 2019) N19-1035
• BERT Post-Training for Review Reading Comprehension and Aspect-based Sentiment Analysis (NAACL 2019) http://arxiv.org/abs/1908.11860
• Exploiting BERT for End-to-End Aspect-based Sentiment Analysis 1910.00883