Private Meeting Summarization Without Performance Loss
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Standard
Private Meeting Summarization Without Performance Loss. / Lee, Seolhwa; Søgaard, Anders.
SIGIR 2023 - Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval. Association for Computing Machinery, Inc., 2023. p. 2282-2286.Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - Private Meeting Summarization Without Performance Loss
AU - Lee, Seolhwa
AU - Søgaard, Anders
N1 - Publisher Copyright: © 2023 Copyright held by the owner/author(s). Publication rights licensed to ACM.
PY - 2023
Y1 - 2023
N2 - Meeting summarization has an enormous business potential, but in addition to being a hard problem, roll-out is challenged by privacy concerns. We explore the problem of meeting summarization under differential privacy constraints and find, to our surprise, that while differential privacy leads to slightly lower performance on in-sample data, differential privacy improves performance when evaluated on unseen meeting types. Since meeting summarization systems will encounter a great variety of meeting types in practical employment scenarios, this observation makes safe meeting summarization seem much more feasible. We perform extensive error analysis and identify potential risks in meeting summarization under differential privacy, including a faithfulness analysis.
AB - Meeting summarization has an enormous business potential, but in addition to being a hard problem, roll-out is challenged by privacy concerns. We explore the problem of meeting summarization under differential privacy constraints and find, to our surprise, that while differential privacy leads to slightly lower performance on in-sample data, differential privacy improves performance when evaluated on unseen meeting types. Since meeting summarization systems will encounter a great variety of meeting types in practical employment scenarios, this observation makes safe meeting summarization seem much more feasible. We perform extensive error analysis and identify potential risks in meeting summarization under differential privacy, including a faithfulness analysis.
KW - Differential Privacy
KW - Meeting Summarization
KW - Text Summarization
U2 - 10.1145/3539618.3592042
DO - 10.1145/3539618.3592042
M3 - Article in proceedings
AN - SCOPUS:85168667313
SP - 2282
EP - 2286
BT - SIGIR 2023 - Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval
PB - Association for Computing Machinery, Inc.
T2 - 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2023
Y2 - 23 July 2023 through 27 July 2023
ER -
ID: 366985696