Linguistic Representations in Multi-task Neural Networks for Ellipsis Resolution

OPEN ARCHIVE

Union Jack
Dannebrog

Linguistic Representations in Multi-task Neural Networks for Ellipsis Resolution

Show simple item record

dc.contributor.author Rønning, Ola
dc.contributor.author Hardt, Daniel
dc.contributor.author Søgaard, Anders
dc.date.accessioned 2018-11-08T15:47:18Z
dc.date.available 2018-11-08T15:47:18Z
dc.date.issued 2018-11-08
dc.identifier.uri http://hdl.handle.net/10398/9684
dc.description.abstract Sluicing resolution is the task of identifying the antecedent to a question ellipsis. Antecedents are often sentential constituents, and previous work has therefore relied on syntactic parsing, together with complex linguistic features. A recent model instead used partial parsing as an auxiliary task in sequential neural network architectures to inject syntactic information. We explore the linguistic information being brought to bear by such networks, both by defining subsets of the data exhibiting relevant linguistic characteristics, and by examining the internal representations of the network. Both perspectives provide evidence for substantial linguistic knowledge being deployed by the neural networks. en_US
dc.format.extent 8 en_US
dc.language eng en_US
dc.title Linguistic Representations in Multi-task Neural Networks for Ellipsis Resolution en_US
dc.type cp en_US
dc.contributor.corporation Copenhagen Business School. CBS en_US
dc.contributor.department Institut for Digitalisering en_US
dc.contributor.departmentshort DIGI en_US
dc.contributor.departmentuk Department of Digitalization en_US
dc.contributor.departmentukshort DIGI en_US
dc.description.notes Paper presented at the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP2018). October 31 - November 4 2018, Brussels, Belgium en_US
dc.publisher.year 2018 en_US


Creative Commons License This work is licensed under a Creative Commons License.

Files Size Format View
Rønning_Hardt_Søgaard.pdf 360.8Kb PDF View/Open Paper

This item appears in the following Collection(s)

Show simple item record