Motivation Existing approaches:Cannbtbeaptyre theiordentof wondslel, edit distance, ..Don't-donsiderthelsyntaaticinformationTobsekpensiveces of a sentenceWepropose aphrase-based similarity model1.Syntacticinformation2.Frequencyinformation3.Lengths ofphrasesSCST@DASFAA2011
Motivation SCST@DASFAA 2011 • Existing approaches: ▪ Word-based, e.g., translation model, edit distance, . ▪ Gram-based, e.g., N-gram, V-gram ▪ All subsequences of a sentence Cannot capture the order of words Don’t consider the syntactic information Too expensive We propose a phrase-based similarity model 1. Syntactic information 2. Frequency information 3. Lengths of phrases
OutlineIntroduction Overview of Our ApproachPhrase-Based Similarity ModelPhraseSelectionExperiments·ConclusionSCST@DASFAA2011
Outline • Introduction • Overview of Our Approach • Phrase-Based Similarity Model • Phrase Selection • Experiments • Conclusion SCST@DASFAA 2011
Problem DefinitionQuery:Query Sentence (English)TranslatorData: ADatabase ofAnswer:ParallelSentencesSentenceswithitstranslationsSentence:English-ChineseSentence:English-ChineseSentences:English-Chinese！SCST@DASFAA2011
Problem Definition SCST@DASFAA 2011 Data: A Database of Parallel Sentences Translator Query: Query Sentence (English) Answer: Sentences with its translations . Sentence1 : English - Chinese Sentence2 : English - Chinese Sentence3 : English - Chinese
Phrase-Based Sentence MatchingPhrase fiPhrase f'1SimilarityPhrase f2Phrase f'2sqModelPhrase fnPhrase f'nOnlineOfflineParallelSentencesPhraseSelectionPhraseDatabaseSCST@DASFAA2011
Phrase-Based Sentence Matching SCST@DASFAA 2011 q Phrase f1 Phrase f2 Phrase fn . s Phrase f ’1 Phrase f ’2 Phrase f ’n . Similarity Model Parallel Sentences Phrase Selection Phrase Database Offline Online
OutlineIntroductionOverview of Our Approach Phrase-Based Similarity ModelPhrase SelectionExperimentsConclusionSCST@DASFAA2011
Outline • Introduction • Overview of Our Approach • Phrase-Based Similarity Model • Phrase Selection • Experiments • Conclusion SCST@DASFAA 2011