We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Semantic Answer Type Prediction using BERT: IAI at the ISWC SMART Task 2020

Abstract: This paper summarizes our participation in the SMART Task of the ISWC 2020 Challenge. A particular question we are interested in answering is how well neural methods, and specifically transformer models, such as BERT, perform on the answer type prediction task compared to traditional approaches. Our main finding is that coarse-grained answer types can be identified effectively with standard text classification methods, with over 95% accuracy, and BERT can bring only marginal improvements. For fine-grained type detection, on the other hand, BERT clearly outperforms previous retrieval-based approaches.
Comments: Published in Proceedings of the SeMantic AnsweR Type prediction task (SMART) at ISWC 2020 Semantic Web Challenge co-located with the 19th International Semantic Web Conference (ISWC 2020). this http URL
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
Journal reference: SMART@ISWC 2020: 10-18
Cite as: arXiv:2109.06714 [cs.CL]
  (or arXiv:2109.06714v1 [cs.CL] for this version)

Submission history

From: Vinay Setty [view email]
[v1] Tue, 14 Sep 2021 14:27:49 GMT (329kb)

Link back to: arXiv, form interface, contact.