Towards Reliable Misinformation Mitigation: Generalization, Uncertainty, and GPT-4
Misinformation poses a critical societal challenge, and current approaches have yet to produce an effective solution. We propose focusing on generalization, uncertainty, and how to leverage recent large language models, in order to create more practical tools to evaluate information veracity in cont...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2023-10 |
---|---|
Hauptverfasser: | , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Pelrine, Kellin Imouza, Anne Thibault, Camille Reksoprodjo, Meilina Gupta, Caleb Christoph, Joel Godbout, Jean-François Rabbany, Reihaneh |
description | Misinformation poses a critical societal challenge, and current approaches have yet to produce an effective solution. We propose focusing on generalization, uncertainty, and how to leverage recent large language models, in order to create more practical tools to evaluate information veracity in contexts where perfect classification is impossible. We first demonstrate that GPT-4 can outperform prior methods in multiple settings and languages. Next, we explore generalization, revealing that GPT-4 and RoBERTa-large exhibit differences in failure modes. Third, we propose techniques to handle uncertainty that can detect impossible examples and strongly improve outcomes. We also discuss results on other language models, temperature, prompting, versioning, explainability, and web retrieval, each one providing practical insights and directions for future research. Finally, we publish the LIAR-New dataset with novel paired English and French misinformation data and Possibility labels that indicate if there is sufficient context for veracity evaluation. Overall, this research lays the groundwork for future tools that can drive real-world progress to combat misinformation. |
format | Article |
fullrecord | <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2819137982</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2819137982</sourcerecordid><originalsourceid>FETCH-proquest_journals_28191379823</originalsourceid><addsrcrecordid>eNqNitEKgjAYRkcQJOU7DLpV0H-a2m2U3QQRdi0rZ0zWv9omUU-fSA_Q1fkO35kQDxiLwzwBmBHf2i6KIlhlkKbMI6dKv7hpLD0JJflFCXqQVmKrzZ07qXFQJ2_jXNNSoDBcyc_oAT3jVRjHJbp3QDk2tDxWYbIg05YrK_wf52S521abffgw-tkL6-pO9waHq4Y8LmKWFTmw_6ov55g_Zg</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2819137982</pqid></control><display><type>article</type><title>Towards Reliable Misinformation Mitigation: Generalization, Uncertainty, and GPT-4</title><source>Free E- Journals</source><creator>Pelrine, Kellin ; Imouza, Anne ; Thibault, Camille ; Reksoprodjo, Meilina ; Gupta, Caleb ; Christoph, Joel ; Godbout, Jean-François ; Rabbany, Reihaneh</creator><creatorcontrib>Pelrine, Kellin ; Imouza, Anne ; Thibault, Camille ; Reksoprodjo, Meilina ; Gupta, Caleb ; Christoph, Joel ; Godbout, Jean-François ; Rabbany, Reihaneh</creatorcontrib><description>Misinformation poses a critical societal challenge, and current approaches have yet to produce an effective solution. We propose focusing on generalization, uncertainty, and how to leverage recent large language models, in order to create more practical tools to evaluate information veracity in contexts where perfect classification is impossible. We first demonstrate that GPT-4 can outperform prior methods in multiple settings and languages. Next, we explore generalization, revealing that GPT-4 and RoBERTa-large exhibit differences in failure modes. Third, we propose techniques to handle uncertainty that can detect impossible examples and strongly improve outcomes. We also discuss results on other language models, temperature, prompting, versioning, explainability, and web retrieval, each one providing practical insights and directions for future research. Finally, we publish the LIAR-New dataset with novel paired English and French misinformation data and Possibility labels that indicate if there is sufficient context for veracity evaluation. Overall, this research lays the groundwork for future tools that can drive real-world progress to combat misinformation.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Classification ; Failure modes ; False information ; Large language models ; Uncertainty</subject><ispartof>arXiv.org, 2023-10</ispartof><rights>2023. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</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>780,784</link.rule.ids></links><search><creatorcontrib>Pelrine, Kellin</creatorcontrib><creatorcontrib>Imouza, Anne</creatorcontrib><creatorcontrib>Thibault, Camille</creatorcontrib><creatorcontrib>Reksoprodjo, Meilina</creatorcontrib><creatorcontrib>Gupta, Caleb</creatorcontrib><creatorcontrib>Christoph, Joel</creatorcontrib><creatorcontrib>Godbout, Jean-François</creatorcontrib><creatorcontrib>Rabbany, Reihaneh</creatorcontrib><title>Towards Reliable Misinformation Mitigation: Generalization, Uncertainty, and GPT-4</title><title>arXiv.org</title><description>Misinformation poses a critical societal challenge, and current approaches have yet to produce an effective solution. We propose focusing on generalization, uncertainty, and how to leverage recent large language models, in order to create more practical tools to evaluate information veracity in contexts where perfect classification is impossible. We first demonstrate that GPT-4 can outperform prior methods in multiple settings and languages. Next, we explore generalization, revealing that GPT-4 and RoBERTa-large exhibit differences in failure modes. Third, we propose techniques to handle uncertainty that can detect impossible examples and strongly improve outcomes. We also discuss results on other language models, temperature, prompting, versioning, explainability, and web retrieval, each one providing practical insights and directions for future research. Finally, we publish the LIAR-New dataset with novel paired English and French misinformation data and Possibility labels that indicate if there is sufficient context for veracity evaluation. Overall, this research lays the groundwork for future tools that can drive real-world progress to combat misinformation.</description><subject>Classification</subject><subject>Failure modes</subject><subject>False information</subject><subject>Large language models</subject><subject>Uncertainty</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNitEKgjAYRkcQJOU7DLpV0H-a2m2U3QQRdi0rZ0zWv9omUU-fSA_Q1fkO35kQDxiLwzwBmBHf2i6KIlhlkKbMI6dKv7hpLD0JJflFCXqQVmKrzZ07qXFQJ2_jXNNSoDBcyc_oAT3jVRjHJbp3QDk2tDxWYbIg05YrK_wf52S521abffgw-tkL6-pO9waHq4Y8LmKWFTmw_6ov55g_Zg</recordid><startdate>20231031</startdate><enddate>20231031</enddate><creator>Pelrine, Kellin</creator><creator>Imouza, Anne</creator><creator>Thibault, Camille</creator><creator>Reksoprodjo, Meilina</creator><creator>Gupta, Caleb</creator><creator>Christoph, Joel</creator><creator>Godbout, Jean-François</creator><creator>Rabbany, Reihaneh</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20231031</creationdate><title>Towards Reliable Misinformation Mitigation: Generalization, Uncertainty, and GPT-4</title><author>Pelrine, Kellin ; Imouza, Anne ; Thibault, Camille ; Reksoprodjo, Meilina ; Gupta, Caleb ; Christoph, Joel ; Godbout, Jean-François ; Rabbany, Reihaneh</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_28191379823</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Classification</topic><topic>Failure modes</topic><topic>False information</topic><topic>Large language models</topic><topic>Uncertainty</topic><toplevel>online_resources</toplevel><creatorcontrib>Pelrine, Kellin</creatorcontrib><creatorcontrib>Imouza, Anne</creatorcontrib><creatorcontrib>Thibault, Camille</creatorcontrib><creatorcontrib>Reksoprodjo, Meilina</creatorcontrib><creatorcontrib>Gupta, Caleb</creatorcontrib><creatorcontrib>Christoph, Joel</creatorcontrib><creatorcontrib>Godbout, Jean-François</creatorcontrib><creatorcontrib>Rabbany, Reihaneh</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Pelrine, Kellin</au><au>Imouza, Anne</au><au>Thibault, Camille</au><au>Reksoprodjo, Meilina</au><au>Gupta, Caleb</au><au>Christoph, Joel</au><au>Godbout, Jean-François</au><au>Rabbany, Reihaneh</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Towards Reliable Misinformation Mitigation: Generalization, Uncertainty, and GPT-4</atitle><jtitle>arXiv.org</jtitle><date>2023-10-31</date><risdate>2023</risdate><eissn>2331-8422</eissn><abstract>Misinformation poses a critical societal challenge, and current approaches have yet to produce an effective solution. We propose focusing on generalization, uncertainty, and how to leverage recent large language models, in order to create more practical tools to evaluate information veracity in contexts where perfect classification is impossible. We first demonstrate that GPT-4 can outperform prior methods in multiple settings and languages. Next, we explore generalization, revealing that GPT-4 and RoBERTa-large exhibit differences in failure modes. Third, we propose techniques to handle uncertainty that can detect impossible examples and strongly improve outcomes. We also discuss results on other language models, temperature, prompting, versioning, explainability, and web retrieval, each one providing practical insights and directions for future research. Finally, we publish the LIAR-New dataset with novel paired English and French misinformation data and Possibility labels that indicate if there is sufficient context for veracity evaluation. Overall, this research lays the groundwork for future tools that can drive real-world progress to combat misinformation.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2023-10 |
issn | 2331-8422 |
language | eng |
recordid | cdi_proquest_journals_2819137982 |
source | Free E- Journals |
subjects | Classification Failure modes False information Large language models Uncertainty |
title | Towards Reliable Misinformation Mitigation: Generalization, Uncertainty, and GPT-4 |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T14%3A12%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Towards%20Reliable%20Misinformation%20Mitigation:%20Generalization,%20Uncertainty,%20and%20GPT-4&rft.jtitle=arXiv.org&rft.au=Pelrine,%20Kellin&rft.date=2023-10-31&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2819137982%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2819137982&rft_id=info:pmid/&rfr_iscdi=true |