Learning to rank answer candidates for automatic resolution of crossword puzzles

Gianni Barlacchi, Massimo Nicosia, Alessandro Moschitti

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

6 Citations (Scopus)

Abstract

In this paper, we study the impact of relational and syntactic representations for an interesting and challenging task: the automatic resolution of crossword puzzles. Automatic solvers are typically based on two answer retrieval modules: (i) a web search engine, e.g., Google, Bing, etc. and (ii) a database (DB) system for accessing previously resolved crossword puzzles. We show that learning to rank models based on relational syntactic structures defined between the clues and the answer can improve both modules above. In particular, our approach accesses the DB using a search engine and reranks its output by modeling paraphrasing. This improves on the MRR of previous system up to 53% in ranking answer candidates and greatly impacts on the resolution accuracy of crossword puzzles up to 15%.

Original languageEnglish
Title of host publicationCoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings
PublisherAssociation for Computational Linguistics (ACL)
Pages39-48
Number of pages10
ISBN (Electronic)9781941643020
Publication statusPublished - 1 Jan 2014
Event18th Conference on Computational Natural Language Learning, CoNLL 2014 - Baltimore, United States
Duration: 26 Jun 201427 Jun 2014

Publication series

NameCoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings

Conference

Conference18th Conference on Computational Natural Language Learning, CoNLL 2014
CountryUnited States
CityBaltimore
Period26/6/1427/6/14

Fingerprint

Syntactics
Search engines
search engine
candidacy
learning
ranking

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Artificial Intelligence
  • Linguistics and Language

Cite this

Barlacchi, G., Nicosia, M., & Moschitti, A. (2014). Learning to rank answer candidates for automatic resolution of crossword puzzles. In CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings (pp. 39-48). (CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings). Association for Computational Linguistics (ACL).

Learning to rank answer candidates for automatic resolution of crossword puzzles. / Barlacchi, Gianni; Nicosia, Massimo; Moschitti, Alessandro.

CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings. Association for Computational Linguistics (ACL), 2014. p. 39-48 (CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings).

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

Barlacchi, G, Nicosia, M & Moschitti, A 2014, Learning to rank answer candidates for automatic resolution of crossword puzzles. in CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings. CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings, Association for Computational Linguistics (ACL), pp. 39-48, 18th Conference on Computational Natural Language Learning, CoNLL 2014, Baltimore, United States, 26/6/14.
Barlacchi G, Nicosia M, Moschitti A. Learning to rank answer candidates for automatic resolution of crossword puzzles. In CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings. Association for Computational Linguistics (ACL). 2014. p. 39-48. (CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings).
Barlacchi, Gianni ; Nicosia, Massimo ; Moschitti, Alessandro. / Learning to rank answer candidates for automatic resolution of crossword puzzles. CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings. Association for Computational Linguistics (ACL), 2014. pp. 39-48 (CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings).
@inproceedings{a0cec3b0c3af4fb59e7b7490b56a705b,
title = "Learning to rank answer candidates for automatic resolution of crossword puzzles",
abstract = "In this paper, we study the impact of relational and syntactic representations for an interesting and challenging task: the automatic resolution of crossword puzzles. Automatic solvers are typically based on two answer retrieval modules: (i) a web search engine, e.g., Google, Bing, etc. and (ii) a database (DB) system for accessing previously resolved crossword puzzles. We show that learning to rank models based on relational syntactic structures defined between the clues and the answer can improve both modules above. In particular, our approach accesses the DB using a search engine and reranks its output by modeling paraphrasing. This improves on the MRR of previous system up to 53{\%} in ranking answer candidates and greatly impacts on the resolution accuracy of crossword puzzles up to 15{\%}.",
author = "Gianni Barlacchi and Massimo Nicosia and Alessandro Moschitti",
year = "2014",
month = "1",
day = "1",
language = "English",
series = "CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings",
publisher = "Association for Computational Linguistics (ACL)",
pages = "39--48",
booktitle = "CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings",

}

TY - GEN

T1 - Learning to rank answer candidates for automatic resolution of crossword puzzles

AU - Barlacchi, Gianni

AU - Nicosia, Massimo

AU - Moschitti, Alessandro

PY - 2014/1/1

Y1 - 2014/1/1

N2 - In this paper, we study the impact of relational and syntactic representations for an interesting and challenging task: the automatic resolution of crossword puzzles. Automatic solvers are typically based on two answer retrieval modules: (i) a web search engine, e.g., Google, Bing, etc. and (ii) a database (DB) system for accessing previously resolved crossword puzzles. We show that learning to rank models based on relational syntactic structures defined between the clues and the answer can improve both modules above. In particular, our approach accesses the DB using a search engine and reranks its output by modeling paraphrasing. This improves on the MRR of previous system up to 53% in ranking answer candidates and greatly impacts on the resolution accuracy of crossword puzzles up to 15%.

AB - In this paper, we study the impact of relational and syntactic representations for an interesting and challenging task: the automatic resolution of crossword puzzles. Automatic solvers are typically based on two answer retrieval modules: (i) a web search engine, e.g., Google, Bing, etc. and (ii) a database (DB) system for accessing previously resolved crossword puzzles. We show that learning to rank models based on relational syntactic structures defined between the clues and the answer can improve both modules above. In particular, our approach accesses the DB using a search engine and reranks its output by modeling paraphrasing. This improves on the MRR of previous system up to 53% in ranking answer candidates and greatly impacts on the resolution accuracy of crossword puzzles up to 15%.

UR - http://www.scopus.com/inward/record.url?scp=85016472126&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85016472126&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85016472126

T3 - CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings

SP - 39

EP - 48

BT - CoNLL 2014 - 18th Conference on Computational Natural Language Learning, Proceedings

PB - Association for Computational Linguistics (ACL)

ER -