ESL Synonym Questions (State of the art)

From ACL Wiki
Revision as of 06:46, 28 February 2008 by Pdturney (talk | contribs)

Jump to: navigation, search
  • ESL = English as a Second Language
  • 50 multiple-choice synonym questions; 4 choices per question
  • each question includes a sentence, providing context for the question
  • ESL questions available from Peter Turney
  • introduced in Turney (2001) as a way of evaluating algorithms for measuring degree of similarity between two words
  • subsequently used by many other researchers

Sample question

Stem: "A rusty nail is not as strong as a clean, new one."
Choices: (a) corroded
(b) black
(c) dirty
(d) painted
Solution: (a) corroded

Table of results

Algorithm Reference for algorithm Reference for experiment Type Correct 95% confidence
Random Random guessing 1 / 4 = 25.00% Random 25.00%
RES Resnik (1995) Jarmasz and Szpakowicz (2003) Hybrid 32.66%
LC Leacock and Chodrow (1998) Jarmasz and Szpakowicz (2003) Lexicon-based 36.00%
LIN Lin (1998) Jarmasz and Szpakowicz (2003) Hybrid 36.00%
JC Jiang and Conrath (1997) Jarmasz and Szpakowicz (2003) Hybrid 36.00%
HSO Hirst and St.-Onge (1998) Jarmasz and Szpakowicz (2003) Lexicon-based 62.00
PMI-IR Turney (2001) Turney (2001) Corpus-based 74.00%
JS Jarmasz and Szpakowicz (2003) Jarmasz and Szpakowicz (2003) Lexicon-based 82.00%

Explanation of table

  • Algorithm = name of algorithm
  • Reference for algorithm = where to find out more about given algorithm
  • Reference for experiment = where to find out more about evaluation of given algorithm with TOEFL questions
  • Type = general type of algorithm: corpus-based, lexicon-based, hybrid
  • Correct = percent of 80 questions that given algorithm answered correctly
  • 95% confidence = confidence interval calculated using Binomial Exact Test
  • table rows sorted in order of increasing percent correct
  • several WordNet-based similarity measures are implemented in Ted Pedersen's WordNet::Similarity package


  • the performance of a corpus-based algorithm depends on the corpus, so the difference in performance between two corpus-based systems may be due to the different corpora, rather than the different algorithms
  • the ESL questions include nouns, verbs, and adjectives, but some of the WordNet-based algorithms were only designed to work with nouns


Jarmasz, M., and Szpakowicz, S. (2003). Roget’s thesaurus and semantic similarity, Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP-03), Borovets, Bulgaria, September, pp. 212-219.

Terra, E., and Clarke, C.L.A. (2003). Frequency estimates for statistical word similarity measures. Proceedings of the Human Language Technology and North American Chapter of Association of Computational Linguistics Conference 2003 (HLT/NAACL 2003), pp. 244–251.

Turney, P.D. (2001). Mining the Web for synonyms: PMI-IR versus LSA on TOEFL. Proceedings of the Twelfth European Conference on Machine Learning (ECML-2001), Freiburg, Germany, pp. 491-502.