Modeling Discriminative Global Inference

Full Text   Posters   Software

Posters:

  • Learning Based Java: A Modeling Language for Discriminative Inference in Learning Based Programs

Authors:

Nick Rizzolo and Dan Roth

Abstract:

Many recent advances in complex domains such as Natural Language Processing
(NLP) have taken a discriminative approach in conjunction with the global
application of structural and domain specific constraints.  We introduce LBJ,
a new modeling language for specifying exact inference systems of this type,
combining ideas from machine learning, optimization, First Order Logic (FOL),
and Object Oriented Programming (OOP).  Expressive constraints are specified
declaratively as arbitrary FOL formulas over functions and objects.  The
language's run-time library translates them to a mathematical programming
representation from which an exact solution is computed.  In addition, the
compiler leverages an existing OOP language: objects and functions are
grounded as the OOP objects and methods that encapsulate the user's
data.

Citation:

N. Rizzolo and D. Roth, Modeling Discriminative Global Inference. ICSC  (2007) pp. 597-604

Bibitem:

@conference{RizzoloRo07,
  author = {N. Rizzolo and D. Roth},
  title = {Modeling Discriminative Global Inference},
  booktitle = {ICSC},
  pages = {597-604},
  month = {9},
  year = {2007},
  address = {Irvine, California},
  publisher = {IEEE},
  url = " http://cogcomp.cs.illinois.edu/papers/RizzoloRo07.pdf",
  funding = {NSF-SoD, LBP},
  projects = {LBP,CCM},
  comment = {Learning Based Java: a Modeling language that facilitates developing of systems with learning and inference componenets.},
}