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

Vis flere oplysninger

Titel: Linguistic Representations in Multi-task Neural Networks for Ellipsis Resolution
Forfatter: Rønning, Ola; Hardt, Daniel; Søgaard, Anders
Resume: 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.
URI: http://hdl.handle.net/10398/9684
Dato: 2018-11-08
Note: Paper presented at the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP2018). October 31 - November 4 2018, Brussels, Belgium

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

Filer Størrelse Format Vis
Rønning_Hardt_Søgaard.pdf 360.8Kb PDF Vis/Åbn Paper

Dette dokument findes i følgende samling(er)

Vis flere oplysninger