Using Integrated Gradients and Constituency Parse Trees to explain Linguistic Acceptability learnt by BERT
This work addresses interpretability for NLP practitioners in tasks like question-answering and machine translation, but it is incremental as it applies existing methods to a specific domain.
The paper tackled the problem of explaining BERT's decision-making in linguistic acceptability classification by using Layer Integrated Gradients and Constituency Parse Trees on the CoLA dataset, finding that correctly classified sentences had 88% to 100% positive attribution scores and misclassified ones had 43% negative scores.
Linguistic Acceptability is the task of determining whether a sentence is grammatical or ungrammatical. It has applications in several use cases like Question-Answering, Natural Language Generation, Neural Machine Translation, where grammatical correctness is crucial. In this paper we aim to understand the decision-making process of BERT (Devlin et al., 2019) in distinguishing between Linguistically Acceptable sentences (LA) and Linguistically Unacceptable sentences (LUA). We leverage Layer Integrated Gradients Attribution Scores (LIG) to explain the Linguistic Acceptability criteria that are learnt by BERT on the Corpus of Linguistic Acceptability (CoLA) (Warstadt et al., 2018) benchmark dataset. Our experiments on 5 categories of sentences lead to the following interesting findings: 1) LIG for LA are significantly smaller in comparison to LUA, 2) There are specific subtrees of the Constituency Parse Tree (CPT) for LA and LUA which contribute larger LIG, 3) Across the different categories of sentences we observed around 88% to 100% of the Correctly classified sentences had positive LIG, indicating a strong positive relationship to the prediction confidence of the model, and 4) Around 43% of the Misclassified sentences had negative LIG, which we believe can become correctly classified sentences if the LIG are parameterized in the loss function of the model.