?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Expressing+Implicit+Semantic+Relations+without+Supervision&rft.creator=Turney%2C+Peter+D.&rft.subject=Language&rft.subject=Computational+Linguistics&rft.subject=Semantics&rft.subject=Machine+Learning&rft.subject=Artificial+Intelligence&rft.description=We+present+an+unsupervised+learning+algorithm+that+mines+large+%0Atext+corpora+for+patterns+that+express+implicit+semantic+relations.+%0AFor+a+given+input+word+pair+X%3AY+with+some+unspecified+semantic+%0Arelations%2C+the+corresponding+output+list+of+patterns+%3CP1%2C...%2CPm%3E++%0Ais+ranked+according+to+how+well+each+pattern+Pi+expresses+the+%0Arelations+between+X+and+Y.+For+example%2C+given+X%3Dostrich+and+%0AY%3Dbird%2C+the+two+highest+ranking+output+patterns+are+%22X+is+the+%0Alargest+Y%22+and+%22Y+such+as+the+X%22.+The+output+patterns+are+intended+%0Ato+be+useful+for+finding+further+pairs+with+the+same+relations%2C+to+%0Asupport+the+construction+of+lexicons%2C+ontologies%2C+and+semantic+%0Anetworks.+The+patterns+are+sorted+by+pertinence%2C+where+the+pertinence+%0Aof+a+pattern+Pi+for+a+word+pair+X%3AY+is+the+expected+relational+%0Asimilarity+between+the+given+pair+and+typical+pairs+for+Pi.+The+%0Aalgorithm+is+empirically+evaluated+on+two+tasks%2C+solving+%0Amultiple-choice+SAT+word+analogy+questions+and+classifying+semantic+%0Arelations+in+noun-modifier+pairs.+On+both+tasks%2C+the+algorithm+%0Aachieves+state-of-the-art+results%2C+performing+significantly+better+%0Athan+several+alternative+pattern+ranking+algorithms%2C+based+on+tf-idf.&rft.date=2006&rft.type=Conference+Paper&rft.type=PeerReviewed&rft.format=application%2Fpdf&rft.identifier=http%3A%2F%2Fcogprints.org%2F5039%2F1%2FNRC-48761.pdf&rft.identifier=++Turney%2C+Peter+D.++(2006)+Expressing+Implicit+Semantic+Relations+without+Supervision.++%5BConference+Paper%5D+++++&rft.relation=http%3A%2F%2Fcogprints.org%2F5039%2F