Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization

Haggai Roitman; Guy Feigenblat; Doron Cohen; Odellia Boni; David Konopnicki

doi:10.1145/3366423.3380009

Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization

Haggai Roitman, Guy Feigenblat, Doron Cohen, Odellia Boni, David Konopnicki

IBM

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

19 Scopus citations

Abstract

We propose Dual-CES - a novel unsupervised, query-focused, multi-document extractive summarizer. Dual-CES builds on top of the Cross Entropy Summarizer (CES) and is designed to better handle the tradeoff between saliency and focus in summarization. To this end, Dual-CES employs a two-step dual-cascade optimization approach with saliency-based pseudo-feedback distillation. Overall, Dual-CES significantly outperforms all other state-of-the-art unsupervised alternatives. Dual-CES is even shown to be able to outperform strong supervised summarizers.

Original language	English
Title of host publication	The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020
Publisher	Association for Computing Machinery, Inc
Pages	2577-2584
Number of pages	8
ISBN (Electronic)	9781450370233
DOIs	https://doi.org/10.1145/3366423.3380009
State	Published - 20 Apr 2020
Externally published	Yes
Event	29th International World Wide Web Conference, WWW 2020 - Taipei, Taiwan, Province of China Duration: 20 Apr 2020 → 24 Apr 2020

Publication series

Name	The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020

Conference

Conference	29th International World Wide Web Conference, WWW 2020
Country/Territory	Taiwan, Province of China
City	Taipei
Period	20/04/20 → 24/04/20

Bibliographical note

Publisher Copyright:
© 2020 ACM.

Access to Document

10.1145/3366423.3380009

Cite this

Roitman, H., Feigenblat, G., Cohen, D., Boni, O., & Konopnicki, D. (2020). Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization. In The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020 (pp. 2577-2584). (The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020). Association for Computing Machinery, Inc. https://doi.org/10.1145/3366423.3380009

Roitman, Haggai ; Feigenblat, Guy ; Cohen, Doron et al. / Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization. The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020. Association for Computing Machinery, Inc, 2020. pp. 2577-2584 (The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020).

@inproceedings{8e8f255000f94544b7cd193bc467b7a4,

title = "Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization",

abstract = "We propose Dual-CES - a novel unsupervised, query-focused, multi-document extractive summarizer. Dual-CES builds on top of the Cross Entropy Summarizer (CES) and is designed to better handle the tradeoff between saliency and focus in summarization. To this end, Dual-CES employs a two-step dual-cascade optimization approach with saliency-based pseudo-feedback distillation. Overall, Dual-CES significantly outperforms all other state-of-the-art unsupervised alternatives. Dual-CES is even shown to be able to outperform strong supervised summarizers.",

author = "Haggai Roitman and Guy Feigenblat and Doron Cohen and Odellia Boni and David Konopnicki",

note = "Publisher Copyright: {\textcopyright} 2020 ACM.; 29th International World Wide Web Conference, WWW 2020 ; Conference date: 20-04-2020 Through 24-04-2020",

year = "2020",

month = apr,

day = "20",

doi = "10.1145/3366423.3380009",

language = "אנגלית",

series = "The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020",

publisher = "Association for Computing Machinery, Inc",

pages = "2577--2584",

booktitle = "The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020",

}

Roitman, H, Feigenblat, G, Cohen, D, Boni, O & Konopnicki, D 2020, Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization. in The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020. The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020, Association for Computing Machinery, Inc, pp. 2577-2584, 29th International World Wide Web Conference, WWW 2020, Taipei, Taiwan, Province of China, 20/04/20. https://doi.org/10.1145/3366423.3380009

Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization. / Roitman, Haggai; Feigenblat, Guy; Cohen, Doron et al.
The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020. Association for Computing Machinery, Inc, 2020. p. 2577-2584 (The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization

AU - Roitman, Haggai

AU - Feigenblat, Guy

AU - Cohen, Doron

AU - Boni, Odellia

AU - Konopnicki, David

PY - 2020/4/20

Y1 - 2020/4/20

N2 - We propose Dual-CES - a novel unsupervised, query-focused, multi-document extractive summarizer. Dual-CES builds on top of the Cross Entropy Summarizer (CES) and is designed to better handle the tradeoff between saliency and focus in summarization. To this end, Dual-CES employs a two-step dual-cascade optimization approach with saliency-based pseudo-feedback distillation. Overall, Dual-CES significantly outperforms all other state-of-the-art unsupervised alternatives. Dual-CES is even shown to be able to outperform strong supervised summarizers.

AB - We propose Dual-CES - a novel unsupervised, query-focused, multi-document extractive summarizer. Dual-CES builds on top of the Cross Entropy Summarizer (CES) and is designed to better handle the tradeoff between saliency and focus in summarization. To this end, Dual-CES employs a two-step dual-cascade optimization approach with saliency-based pseudo-feedback distillation. Overall, Dual-CES significantly outperforms all other state-of-the-art unsupervised alternatives. Dual-CES is even shown to be able to outperform strong supervised summarizers.

UR - https://www.scopus.com/pages/publications/85086592620

U2 - 10.1145/3366423.3380009

DO - 10.1145/3366423.3380009

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85086592620

T3 - The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020

SP - 2577

EP - 2584

BT - The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020

PB - Association for Computing Machinery, Inc

T2 - 29th International World Wide Web Conference, WWW 2020

Y2 - 20 April 2020 through 24 April 2020

ER -

Roitman H, Feigenblat G, Cohen D, Boni O, Konopnicki D. Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization. In The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020. Association for Computing Machinery, Inc. 2020. p. 2577-2584. (The Web Conference 2020 - Proceedings of the World Wide Web Conference, WWW 2020). doi: 10.1145/3366423.3380009

Unsupervised Dual-Cascade Learning with Pseudo-Feedback Distillation for Query-Focused Extractive Summarization

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this