TY - GEN
T1 - Efficient Attentions for Long Document Summarization
AU - Huang, Luyang
AU - Cao, Shuyang
AU - Parulian, Nikolaus
AU - Ji, Heng
AU - Wang, Lu
N1 - Funding Information:
This research is supported in part by Oracle for Research Cloud Credits, National Science Foundation through Grant IIS-1813341, and by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via contract # FA8650-17-C-9116. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of ODNI, IARPA, or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation therein. We thank three anonymous reviewers for their valuable suggestions and comments.
Publisher Copyright:
© 2021 Association for Computational Linguistics.
PY - 2021
Y1 - 2021
N2 - The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose HEPOS, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with HEPOS, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GOVREPORT, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.
AB - The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose HEPOS, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with HEPOS, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GOVREPORT, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.
UR - http://www.scopus.com/inward/record.url?scp=85115367009&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85115367009&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85115367009
T3 - NAACL-HLT 2021 - 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference
SP - 1419
EP - 1436
BT - NAACL-HLT 2021 - 2021 Conference of the North American Chapter of the Association for Computational Linguistics
PB - Association for Computational Linguistics (ACL)
T2 - 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2021
Y2 - 6 June 2021 through 11 June 2021
ER -