BERT-kNN: Adding a kNN Search Component to Pretrained Language Models for Better QA
This work addresses fact recall issues in QA for users needing accurate and up-to-date information, representing an incremental improvement over existing methods.
The paper tackles the problem of improving open-domain question answering by combining BERT with a kNN search component, resulting in BERT-kNN outperforming BERT on cloze-style QA by large margins without further training, excelling for rare facts and handling recent events.
Khandelwal et al. (2020) use a k-nearest-neighbor (kNN) component to improve language model performance. We show that this idea is beneficial for open-domain question answering (QA). To improve the recall of facts encountered during training, we combine BERT (Devlin et al., 2019) with a traditional information retrieval step (IR) and a kNN search over a large datastore of an embedded text collection. Our contributions are as follows: i) BERT-kNN outperforms BERT on cloze-style QA by large margins without any further training. ii) We show that BERT often identifies the correct response category (e.g., US city), but only kNN recovers the factually correct answer (e.g., "Miami"). iii) Compared to BERT, BERT-kNN excels for rare facts. iv) BERT-kNN can easily handle facts not covered by BERT's training set, e.g., recent events.