TY - JOUR
T1 - Automatically Summarizing Evidence from Clinical Trials
T2 - 17th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2023
AU - Ramprasad, Sanjana
AU - McInerney, Jered
AU - Marshall, Iain J.
AU - Wallace, Byron C.
N1 - Funding Information:
This work was supported in part by the National Institutes of Health (NIH) under award R01LM012086, and by the National Science Foundation (NSF) awards 1901117 and 2211954. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the NSF.
Publisher Copyright:
© 2023 Association for Computational Linguistics.
PY - 2023/5
Y1 - 2023/5
N2 - We present TrialsSummarizer, a system that aims to automatically summarize evidence presented in the set of randomized controlled trials most relevant to a given query. Building on prior work (Marshall et al., 2020), the system retrieves trial publications matching a query specifying a combination of condition, intervention(s), and outcome(s), and ranks these according to sample size and estimated study quality. The top-k such studies are passed through a neural multi-document summarization system, yielding a synopsis of these trials. We consider two architectures: A standard sequence-to-sequence model based on BART (Lewis et al., 2019), and a multi-headed architecture intended to provide greater transparency to end-users. Both models produce fluent and relevant summaries of evidence retrieved for queries, but their tendency to introduce unsupported statements render them inappropriate for use in this domain at present. The proposed architecture may help users verify outputs allowing users to trace generated tokens back to inputs. The demonstration video is available at: https://vimeo.com/735605060 The prototype, source code, and model weights are available at: https://sanjanaramprasad.github.io/trials-summarizer/.
AB - We present TrialsSummarizer, a system that aims to automatically summarize evidence presented in the set of randomized controlled trials most relevant to a given query. Building on prior work (Marshall et al., 2020), the system retrieves trial publications matching a query specifying a combination of condition, intervention(s), and outcome(s), and ranks these according to sample size and estimated study quality. The top-k such studies are passed through a neural multi-document summarization system, yielding a synopsis of these trials. We consider two architectures: A standard sequence-to-sequence model based on BART (Lewis et al., 2019), and a multi-headed architecture intended to provide greater transparency to end-users. Both models produce fluent and relevant summaries of evidence retrieved for queries, but their tendency to introduce unsupported statements render them inappropriate for use in this domain at present. The proposed architecture may help users verify outputs allowing users to trace generated tokens back to inputs. The demonstration video is available at: https://vimeo.com/735605060 The prototype, source code, and model weights are available at: https://sanjanaramprasad.github.io/trials-summarizer/.
UR - http://www.scopus.com/inward/record.url?scp=85159857337&partnerID=8YFLogxK
U2 - 10.18653/v1/2023.eacl-demo.27
DO - 10.18653/v1/2023.eacl-demo.27
M3 - Conference paper
AN - SCOPUS:85159857337
SP - 236
EP - 247
JO - Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations
JF - Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations
Y2 - 2 May 2023 through 6 May 2023
ER -