BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning hard chart constraints for efficient context-free parsing
  - Brian Roark - Oregon Health and Science University
DTSTART:20111006T110000Z
DTEND:20111006T120000Z
UID:TALK33175@talks.cam.ac.uk
CONTACT:Thomas Lippincott
DESCRIPTION: In this talk\, I'll present some recent work in learning hard
  constraints for cells within a context-free parsing chart\, to reduce par
 sing time. Each cell in the chart represents one of the O(n^2) substrings 
 of the input string\, and characteristics of each substring can be used to
  decide how much work to do in the associated chart cell. I'll discuss fin
 ite-state models for tagging chart constraints on words\, including method
 s for bounding the worst-case complexity of the parsing pipeline to quadra
 tic or sub-quadratic in the length of the string. Empirical results will b
 e presented for English and Chinese\, achieved by constraining various hig
 h accuracy parsers.  Finally\, I will present a generalization of these fi
 nite-state approaches that performs a quadratic number of classifications 
 (one for each substring) to produce further (finer) constraints on the amo
 unt of processing within each cell. This latter approach has the nice prop
 erty of being trained on maximum likelihood parses\, rather than reference
  parses\, making for a straightforward method for tuning parsing efficienc
 y to new tasks and domains.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR