Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint

Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wu, Heng-Yi, Zhang, Jingqing, Ive, Julia, Li, Tong, Gupta, Vibhor, Chen, Bingyuan, Guo, Yike
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Wu, Heng-Yi
Zhang, Jingqing
Ive, Julia
Li, Tong
Gupta, Vibhor
Chen, Bingyuan
Guo, Yike
description Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the data paucity, data sparsity and inability of the state-of-the-art natural language generation models (including T5, PEGASUS and GPT-Neo) to produce accurate and reliable outputs. In this paper, we propose a novel table-to-text approach and tackle these problems with a novel two-step architecture which is enhanced by auto-correction, copy mechanism and synthetic data augmentation. The study shows that the proposed approach selects salient biomedical entities and values from structured data with improved precision (up to 0.13 absolute increase) of copying the tabular values to generate coherent and accurate text for assay validation reports and toxicology reports. Moreover, we also demonstrate a light-weight adaptation of the proposed system to new datasets by fine-tuning with as little as 40\% training examples. The outputs of our model are validated by human experts in the Human-in-the-Loop scenario.
doi_str_mv 10.48550/arxiv.2205.12368
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2205_12368</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2205_12368</sourcerecordid><originalsourceid>FETCH-LOGICAL-a678-81065bc6343f3232bbc63b0f8afdfbbcb7cc0ff96ae9ddc841be48db2df4f1653</originalsourceid><addsrcrecordid>eNotj7FOwzAURb0woMIHMOEfcHDsxHVHFKBFSsXQ7NGz_axaSp3IcaH9e9rCdO9ZjnQIeSp5Uem65i-QTuG7EILXRSmk0vcEt-iChYHubMCYgw-WdmAGZHlkHZ4yXWPEBDmMkf6EvKeb4wEiC5HlPbJ2HCd6jA4TvSB9gwx0N0GaQz7TZoxzThBifiB3HoYZH_93QbqP967ZsPZr_dm8tgzUUjNdclUbq2QlvRRSmOs33Gvwzl_ALK3l3q8U4Mo5q6vSYKWdEc5XvlS1XJDnP-2ts59SOEA699fe_tYrfwHDK1J2</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint</title><source>arXiv.org</source><creator>Wu, Heng-Yi ; Zhang, Jingqing ; Ive, Julia ; Li, Tong ; Gupta, Vibhor ; Chen, Bingyuan ; Guo, Yike</creator><creatorcontrib>Wu, Heng-Yi ; Zhang, Jingqing ; Ive, Julia ; Li, Tong ; Gupta, Vibhor ; Chen, Bingyuan ; Guo, Yike</creatorcontrib><description>Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the data paucity, data sparsity and inability of the state-of-the-art natural language generation models (including T5, PEGASUS and GPT-Neo) to produce accurate and reliable outputs. In this paper, we propose a novel table-to-text approach and tackle these problems with a novel two-step architecture which is enhanced by auto-correction, copy mechanism and synthetic data augmentation. The study shows that the proposed approach selects salient biomedical entities and values from structured data with improved precision (up to 0.13 absolute increase) of copying the tabular values to generate coherent and accurate text for assay validation reports and toxicology reports. Moreover, we also demonstrate a light-weight adaptation of the proposed system to new datasets by fine-tuning with as little as 40\% training examples. The outputs of our model are validated by human experts in the Human-in-the-Loop scenario.</description><identifier>DOI: 10.48550/arxiv.2205.12368</identifier><language>eng</language><subject>Computer Science - Computation and Language</subject><creationdate>2022-05</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2205.12368$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2205.12368$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Wu, Heng-Yi</creatorcontrib><creatorcontrib>Zhang, Jingqing</creatorcontrib><creatorcontrib>Ive, Julia</creatorcontrib><creatorcontrib>Li, Tong</creatorcontrib><creatorcontrib>Gupta, Vibhor</creatorcontrib><creatorcontrib>Chen, Bingyuan</creatorcontrib><creatorcontrib>Guo, Yike</creatorcontrib><title>Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint</title><description>Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the data paucity, data sparsity and inability of the state-of-the-art natural language generation models (including T5, PEGASUS and GPT-Neo) to produce accurate and reliable outputs. In this paper, we propose a novel table-to-text approach and tackle these problems with a novel two-step architecture which is enhanced by auto-correction, copy mechanism and synthetic data augmentation. The study shows that the proposed approach selects salient biomedical entities and values from structured data with improved precision (up to 0.13 absolute increase) of copying the tabular values to generate coherent and accurate text for assay validation reports and toxicology reports. Moreover, we also demonstrate a light-weight adaptation of the proposed system to new datasets by fine-tuning with as little as 40\% training examples. The outputs of our model are validated by human experts in the Human-in-the-Loop scenario.</description><subject>Computer Science - Computation and Language</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj7FOwzAURb0woMIHMOEfcHDsxHVHFKBFSsXQ7NGz_axaSp3IcaH9e9rCdO9ZjnQIeSp5Uem65i-QTuG7EILXRSmk0vcEt-iChYHubMCYgw-WdmAGZHlkHZ4yXWPEBDmMkf6EvKeb4wEiC5HlPbJ2HCd6jA4TvSB9gwx0N0GaQz7TZoxzThBifiB3HoYZH_93QbqP967ZsPZr_dm8tgzUUjNdclUbq2QlvRRSmOs33Gvwzl_ALK3l3q8U4Mo5q6vSYKWdEc5XvlS1XJDnP-2ts59SOEA699fe_tYrfwHDK1J2</recordid><startdate>20220524</startdate><enddate>20220524</enddate><creator>Wu, Heng-Yi</creator><creator>Zhang, Jingqing</creator><creator>Ive, Julia</creator><creator>Li, Tong</creator><creator>Gupta, Vibhor</creator><creator>Chen, Bingyuan</creator><creator>Guo, Yike</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20220524</creationdate><title>Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint</title><author>Wu, Heng-Yi ; Zhang, Jingqing ; Ive, Julia ; Li, Tong ; Gupta, Vibhor ; Chen, Bingyuan ; Guo, Yike</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a678-81065bc6343f3232bbc63b0f8afdfbbcb7cc0ff96ae9ddc841be48db2df4f1653</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Computer Science - Computation and Language</topic><toplevel>online_resources</toplevel><creatorcontrib>Wu, Heng-Yi</creatorcontrib><creatorcontrib>Zhang, Jingqing</creatorcontrib><creatorcontrib>Ive, Julia</creatorcontrib><creatorcontrib>Li, Tong</creatorcontrib><creatorcontrib>Gupta, Vibhor</creatorcontrib><creatorcontrib>Chen, Bingyuan</creatorcontrib><creatorcontrib>Guo, Yike</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Wu, Heng-Yi</au><au>Zhang, Jingqing</au><au>Ive, Julia</au><au>Li, Tong</au><au>Gupta, Vibhor</au><au>Chen, Bingyuan</au><au>Guo, Yike</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint</atitle><date>2022-05-24</date><risdate>2022</risdate><abstract>Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the data paucity, data sparsity and inability of the state-of-the-art natural language generation models (including T5, PEGASUS and GPT-Neo) to produce accurate and reliable outputs. In this paper, we propose a novel table-to-text approach and tackle these problems with a novel two-step architecture which is enhanced by auto-correction, copy mechanism and synthetic data augmentation. The study shows that the proposed approach selects salient biomedical entities and values from structured data with improved precision (up to 0.13 absolute increase) of copying the tabular values to generate coherent and accurate text for assay validation reports and toxicology reports. Moreover, we also demonstrate a light-weight adaptation of the proposed system to new datasets by fine-tuning with as little as 40\% training examples. The outputs of our model are validated by human experts in the Human-in-the-Loop scenario.</abstract><doi>10.48550/arxiv.2205.12368</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2205.12368
ispartof
issn
language eng
recordid cdi_arxiv_primary_2205_12368
source arXiv.org
subjects Computer Science - Computation and Language
title Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-05T10%3A43%3A14IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Medical%20Scientific%20Table-to-Text%20Generation%20with%20Human-in-the-Loop%20under%20the%20Data%20Sparsity%20Constraint&rft.au=Wu,%20Heng-Yi&rft.date=2022-05-24&rft_id=info:doi/10.48550/arxiv.2205.12368&rft_dat=%3Carxiv_GOX%3E2205_12368%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true