Show simple item record

dc.contributor.advisorSurdeanu, Mihai
dc.contributor.authorMithun, Mitch Paul
dc.creatorMithun, Mitch Paul
dc.date.accessioned2022-05-19T18:59:50Z
dc.date.available2022-05-19T18:59:50Z
dc.date.issued2022
dc.identifier.citationMithun, Mitch Paul. (2022). Knowledge Distillation as a Solution for Domain Transfer (Doctoral dissertation, University of Arizona, Tucson, USA).
dc.identifier.urihttp://hdl.handle.net/10150/664293
dc.description.abstractWhile neural networks produce state-of-the- art performance in several NLP tasks, they generally depend heavily on lexicalized information, which transfer poorly between domains. We present a combination of two strategies to mitigate this dependence on lexicalized information in fact verification tasks. We present a data distillation technique for delexicalization, which we then combine with a model distillation method to prevent aggressive data distillation. We show that by using our solution, not only does the performance of an ex- isting state-of-the-art model remain at par with that of the model trained on a fully lexicalized data, but it also performs better than it when tested out of domain. We show that the technique we present encourages models to extract transferable facts from a given fact verification dataset.
dc.language.isoen
dc.publisherThe University of Arizona.
dc.rightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/
dc.subjectdelexicalization
dc.subjectdomain adaptation
dc.subjectKnowledge Distillation
dc.subjectnatural language processing
dc.titleKnowledge Distillation as a Solution for Domain Transfer
dc.typetext
dc.typeElectronic Dissertation
thesis.degree.grantorUniversity of Arizona
thesis.degree.leveldoctoral
dc.contributor.committeememberBarnard, Jacobus
dc.contributor.committeememberMorrison, Clayton
dc.contributor.committeememberBrown, Sue
thesis.degree.disciplineGraduate College
thesis.degree.disciplineComputer Science
thesis.degree.namePh.D.
refterms.dateFOA2022-05-19T18:59:50Z


Files in this item

Thumbnail
Name:
azu_etd_19496_sip1_m.pdf
Size:
1.933Mb
Format:
PDF

This item appears in the following Collection(s)

Show simple item record