Incremental Reading for Question Answering

Incremental Reading for Question Answering

Any system which performs goal-directed continual learning must not only learn incrementally but process and absorb information incrementally. Such a system also has to understand when its goals have been achieved. In this paper, we consider these issues in the context of question answering. Current...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Abnar, Samira, Bedrax-weiss, Tania, Kwiatkowski, Tom, Cohen, William W
Format: Artikel
Sprache:eng
Schlagworte:
Computer Science - Artificial Intelligence
Computer Science - Computation and Language
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 Abnar, Samira
Bedrax-weiss, Tania
Kwiatkowski, Tom
Cohen, William W
description Any system which performs goal-directed continual learning must not only learn incrementally but process and absorb information incrementally. Such a system also has to understand when its goals have been achieved. In this paper, we consider these issues in the context of question answering. Current state-of-the-art question answering models reason over an entire passage, not incrementally. As we will show, naive approaches to incremental reading, such as restriction to unidirectional language models in the model, perform poorly. We present extensions to the DocQA [2] model to allow incremental reading without loss of accuracy. The model also jointly learns to provide the best answer given the text that is seen so far and predict whether this best-so-far answer is sufficient.
doi_str_mv 10.48550/arxiv.1901.04936
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_1901_04936</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>1901_04936</sourcerecordid><originalsourceid>FETCH-LOGICAL-a676-bc3f944470728b234fe7fbe4f80f9d15eb77faec2c3f54fc4c6078114ee8b4c3</originalsourceid><addsrcrecordid>eNotzkkLwjAUBOBcPIj6AzzZs9CaNK9NehRxA0Fc7iWJ70lBo6Su_971NDAMw8dYV_AEdJbxgQmP6paIgouEQyHzJuvPvQt4RH8xh2iNZlf5fUSnEK2uWF-qk4-Gvr5jeNdt1iBzqLHzzxbbTMbb0SxeLKfz0XARm1zlsXWSCgBQXKXaphIIFVkE0pyKncjQKkUGXfreZUAOXM6VFgIQtQUnW6z3e_1ay3OojiY8y4-5_JrlC2k3PBw</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Incremental Reading for Question Answering</title><source>arXiv.org</source><creator>Abnar, Samira ; Bedrax-weiss, Tania ; Kwiatkowski, Tom ; Cohen, William W</creator><creatorcontrib>Abnar, Samira ; Bedrax-weiss, Tania ; Kwiatkowski, Tom ; Cohen, William W</creatorcontrib><description>Any system which performs goal-directed continual learning must not only learn incrementally but process and absorb information incrementally. Such a system also has to understand when its goals have been achieved. In this paper, we consider these issues in the context of question answering. Current state-of-the-art question answering models reason over an entire passage, not incrementally. As we will show, naive approaches to incremental reading, such as restriction to unidirectional language models in the model, perform poorly. We present extensions to the DocQA [2] model to allow incremental reading without loss of accuracy. The model also jointly learns to provide the best answer given the text that is seen so far and predict whether this best-so-far answer is sufficient.</description><identifier>DOI: 10.48550/arxiv.1901.04936</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Computation and Language</subject><creationdate>2019-01</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</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/1901.04936$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.1901.04936$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Abnar, Samira</creatorcontrib><creatorcontrib>Bedrax-weiss, Tania</creatorcontrib><creatorcontrib>Kwiatkowski, Tom</creatorcontrib><creatorcontrib>Cohen, William W</creatorcontrib><title>Incremental Reading for Question Answering</title><description>Any system which performs goal-directed continual learning must not only learn incrementally but process and absorb information incrementally. Such a system also has to understand when its goals have been achieved. In this paper, we consider these issues in the context of question answering. Current state-of-the-art question answering models reason over an entire passage, not incrementally. As we will show, naive approaches to incremental reading, such as restriction to unidirectional language models in the model, perform poorly. We present extensions to the DocQA [2] model to allow incremental reading without loss of accuracy. The model also jointly learns to provide the best answer given the text that is seen so far and predict whether this best-so-far answer is sufficient.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Computation and Language</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotzkkLwjAUBOBcPIj6AzzZs9CaNK9NehRxA0Fc7iWJ70lBo6Su_971NDAMw8dYV_AEdJbxgQmP6paIgouEQyHzJuvPvQt4RH8xh2iNZlf5fUSnEK2uWF-qk4-Gvr5jeNdt1iBzqLHzzxbbTMbb0SxeLKfz0XARm1zlsXWSCgBQXKXaphIIFVkE0pyKncjQKkUGXfreZUAOXM6VFgIQtQUnW6z3e_1ay3OojiY8y4-5_JrlC2k3PBw</recordid><startdate>20190115</startdate><enddate>20190115</enddate><creator>Abnar, Samira</creator><creator>Bedrax-weiss, Tania</creator><creator>Kwiatkowski, Tom</creator><creator>Cohen, William W</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20190115</creationdate><title>Incremental Reading for Question Answering</title><author>Abnar, Samira ; Bedrax-weiss, Tania ; Kwiatkowski, Tom ; Cohen, William W</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a676-bc3f944470728b234fe7fbe4f80f9d15eb77faec2c3f54fc4c6078114ee8b4c3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Computation and Language</topic><toplevel>online_resources</toplevel><creatorcontrib>Abnar, Samira</creatorcontrib><creatorcontrib>Bedrax-weiss, Tania</creatorcontrib><creatorcontrib>Kwiatkowski, Tom</creatorcontrib><creatorcontrib>Cohen, William W</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Abnar, Samira</au><au>Bedrax-weiss, Tania</au><au>Kwiatkowski, Tom</au><au>Cohen, William W</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Incremental Reading for Question Answering</atitle><date>2019-01-15</date><risdate>2019</risdate><abstract>Any system which performs goal-directed continual learning must not only learn incrementally but process and absorb information incrementally. Such a system also has to understand when its goals have been achieved. In this paper, we consider these issues in the context of question answering. Current state-of-the-art question answering models reason over an entire passage, not incrementally. As we will show, naive approaches to incremental reading, such as restriction to unidirectional language models in the model, perform poorly. We present extensions to the DocQA [2] model to allow incremental reading without loss of accuracy. The model also jointly learns to provide the best answer given the text that is seen so far and predict whether this best-so-far answer is sufficient.</abstract><doi>10.48550/arxiv.1901.04936</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.1901.04936
ispartof
issn
language eng
recordid cdi_arxiv_primary_1901_04936
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Computation and Language
title Incremental Reading for Question Answering
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-12T10%3A10%3A58IST&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=Incremental%20Reading%20for%20Question%20Answering&rft.au=Abnar,%20Samira&rft.date=2019-01-15&rft_id=info:doi/10.48550/arxiv.1901.04936&rft_dat=%3Carxiv_GOX%3E1901_04936%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