UA CSC 620 - Advanced Topics in Natural Language Processing

Unformatted text preview:

C SC 620 Advanced Topics in Natural Language ProcessingWordNet Browser SoftwareSlide 3wnconnect SoftwareSlide 5PowerPoint PresentationProlog as a Database Query LanguageSlide 8Class ExercisePossible Software ProjectAutoantonymsAntagonymsC SC 620Advanced Topics in Natural Language ProcessingLecture Notes 31/22/04WordNet Browser SoftwareWordNet Browser Software•For most platforms, see the Princeton website:–http://www.cogsci.princeton.edu/~wn/•For MacOS X (the one I’ve been using), see:–http://wordnet.sourceforge.net/wnconnect Software•http://linguistics.arizona.edu/~sandiway/wnconnect/wnconnect Software•Does a breadth-first search to find shortest or all connections between two words–MacOS X version uses neato (from the freely available Graphviz package) to render graphs•TTY-interface version now available for the Windows platform –(courtesy of TszYan Sandy Chow)–Uses SWI-Prolog •Free download from www.swi-prolog.org–Usage (Prolog syntax):•?- connect(happy,sad).Prolog as a Database Query Language•Can use other tools, of course …•Prolog as a logic-based database query language:–Bust as a verb belongs to 5 different synsets, i.e. has five different senses: •?- findall(S),s2(bust,v,_,S,_),L).L = [200266721,201083468,201083844,201246161,201961102] •Three of these have the same file number 35 (verb.contact): •?- findall((S,F),(s2(bust,v,_,S,_),file(S,F)),L).•L = [(200266721,30),(201083468,35), (201083844,35),(201246161,35),(201961102,41)]Prolog as a Database Query Language–The glosses corresponding to these three senses are:• ?- findall((S,G),(s2(bust,v,I,S,_),file(S,35),g(I,G)),L).•L = –[(201083468,'(go to pieces; "The lawn mower finally broke"; "The gears wore out"; "The old chair finally fell apart completely")'),–(201083844,'(ruin completely; "He busted my radio!")'),–(201246161,'(separate or cause to separate abruptly; "The rope snapped"; "tear the paper")')] –The relation common file number will conflate these three senses.Class Exercise•What have (specific senses) of the following nouns in common?–Umbrella–Saucepan–Baseball bat–Carpet beater•But do not share with:–Giraffe–Pretzel–HomeworkPossible Software Project•Verb autoantonymy “Verbs that express both one meaning and an opposing meaning”–Example: dust–American Heritage Dictionary:1.To remove dust …•dust the furniture2.To apply dust …•dust the cookies with sugarAutoantonyms•Locate?–There is a website with a manually-compiled list of autoantonyms somewhere … •Possibly, it is this page on antagonyms:–http://www-personal.umich.edu/~cellis/antagonym.htmlAntagonyms•Examples:–Buckle: to hold together (e.g. buckle your belt) vs. to fall apart (e.g., buckle under pressure) {AQ} Bull: A solemn edict or mandate vs. Nonsense or worthless information (3) {M} –Chuff: Elated vs. Unhappy (hinted at in 1) {M} –Cite, Citation: For doing good (such as military gallantry) vs. for doing bad (such as from a traffic policeman) (1) –Cleave: To adhere tightly vs. To cut apart (1) {A} –Clip: to attach vs. to cut off (1) {AH}{AS} –Cool: positive sense (cool web-sites) vs. negative sense(cool reception).


View Full Document

UA CSC 620 - Advanced Topics in Natural Language Processing

Download Advanced Topics in Natural Language Processing
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Advanced Topics in Natural Language Processing and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Advanced Topics in Natural Language Processing 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?