Automatic feature engineering for Italian question answering systems

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we propose automatic feature engineering for Italian QA systems. Our approach only requires a shallow syntactic rep- resentation of the questions and the answer passages. We apply Support Vector Machines using tree kernels to such trees for automatically gener- ating relational syntactic patters, which significantly improve on BM25 retrieval models.

Original languageEnglish
Title of host publicationCEUR Workshop Proceedings
PublisherCEUR-WS
Volume1404
Publication statusPublished - 2015
Event6th Italian Information Retrieval Workshop, IIR 2015 - Cagliari, Italy
Duration: 25 May 201526 May 2015

Other

Other6th Italian Information Retrieval Workshop, IIR 2015
CountryItaly
CityCagliari
Period25/5/1526/5/15

    Fingerprint

Keywords

  • KernelMeth-ods
  • Question Answering
  • Support Vector Machines

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Uva, A., & Moschitti, A. (2015). Automatic feature engineering for Italian question answering systems. In CEUR Workshop Proceedings (Vol. 1404). CEUR-WS.