50 likes | 63 Views
This presentation covers various methods for word association, including independence tests, mutual information, Dice and Jaccard measures. It also explores the effectiveness of different measures, such as log-likelihood association, in finding collocations and terminology. Additionally, the presentation discusses word similarity, scalar words, orientation, opinion detection, ontologies and lexical databases, word and phrase alignment, sense disambiguation, and bioinformatics terminology and classification.
E N D
Topics for Presentations Vasileios Hatzivassiloglou University of Texas at Dallas
Presented by me • Word Association (Independence tests; mutual information; Dice and Jaccard measures) • Word Association: Comparisons • Smadja, McKeown, and Hatzivassiloglou on the effectiveness of different measures • Dunning on log-likelihood association measure • Collocations and Terminology • Smadja on finding collocations • Daille on finding terms via association • Justeson & Katz on terms via recurrence
Presented by you • Word similarity: early work (Shütze 1992; Pereira et al., ACL-93; Hatzivassiloglou and McKeown, ACL-93) • Word similarity: IR and comparison (Smeaton SIGIR-96; Lee ACL-99; Terra and Clarke, NAACL-03) • Scalar words • Elhadad 1991 on generation • Raskin and Niremburg on ontologies • Marcu and Hirst ACL-95 on inference
Scalar implicature (Sedivy et al. 1999; Noveck 2001; Papafragou and Musolino 2003) • Orientation • Hatzivassiloglou and McKeown, ACL 97 on polarity of adjectives • Wiebe 2001 on subjective words • Riloff, Wiebe, and Cardie on subjective words via bootstrapping • Opinion detection • Two document-based approaches (product reviews) • Sentence based (opinion Q&A) • Ontologies and Lexical Databases • WordNet, CYC • SENSUS and expanding it automatically (Knight 1993) • Ontology creation • Two papers on text mining for ontologies • Measuring semantic distance in an ontology (Resnik 1995)
Word and Phrase Alignment • Bilingual dictionaries (Klavans and Chodorow 1991) • Alignment (Smadja et al. 1996 in part; Fung 1996) • Paraphrase detection (Barzilay 2002) • Sense disambiguation • Early work by Gale et al., Brown et al. • Yarowsky: One sense per discourse/collocation, extensions • Bioinformatics: Terminology and Classification • Term detection via morphology (Krauthammer 2002) • Disambiguation / classification: Hatzivassiloglou et al. 2001 • Acronyms (Yu et al. 2002) • Bioinformatics: Functional Relationships • Work by Craven, Shütze, and Altman (2 or 3 papers)