MULTI-SPAN QUESTION ANSWERING USING SPAN-IMAGE NETWORK

Tarik Arici, Hayreddin Ceker, Ismail Baha Tutar

Keywords: BERT, deep learning, multi-span answer, question-answering, SQuAD, transformers

Abstract: Question-answering (QA) models aim to find an answer given a question and con- text. Language models like BERT are used to associate question and context to find an answer span. Prior art on QA focuses on finding the best answer. There is a need for multi-span QA models to output the top-K likely answers to questions such as "Which companies Elon Musk started?" or "What factors cause global warming?" In this work, we introduce Span-Image architecture that can learn to identify multiple answers in a context for a given question. This architecture can incorporate prior information about the span length distribution or valid span patterns (e.g., end index has to be larger than start index), thus eliminating the need for post-processing. Span-Image architecture outperforms the state-of-the-art in top-K answer accuracy on SQuAD dataset and in multi-span answer accuracy on an Amazon internal dataset.

One-sentence Summary: We build multi-span question-answering models to output the top-N likely answers to questions instead of one answer using SQuAD dataset and Amazon internal dataset.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Reviewed Version (pdf): https://openreview.net/references/pdf?id=QxGaN_qJM6