Abstractive Summarization of Spoken andWritten Instructions with BERT

Summarization of speech is a difficult problem due to the spontaneity of the flow, disfluencies, and other issues that are not usually encountered in written texts . We use transfer learning and pretrain the model on a few large cross-domain datasets in both written and spoken English . The model beats current SOTA when applied to WikiHow articles that vary widely in style and topic, while showing no performance regression on the canonical CNN/DailyMail dataset . Due to the high generalizability of the model, it has great potential to improve accessibility and discoverability of internet content . We envision this integrated as a feature in intelligent virtual assistants, enabling them to . summarize both written . and spoken instructional content upon request . We also do preprocessing of transcripts to restore sentence segmentation and punctuation in the output of an ASR system . The results are evaluated with ROUGE and Content-F1 scoring for the How2 and WikiHow datasets. We engage human judges to score a set of summaries randomly selected from a dataset curated from HowTo100M and YouTube. Based on blind evaluation, we achieve a level of textual fluency and utility close to that of . summaries

Links: PDF - Abstract

Code :

https://github.com/alebryvas/berk266
https://github.com/nlpyang/PreSumm

Keywords : model - spoken - content - written - wikihow -

Leave a Reply

Your email address will not be published. Required fields are marked *