There is a sign-up schedule for this event that can be found at

http://www.cs.utexas.edu/department/webevent/utcs/events/cgi/list_event

s.cgi

Type of Talk: UTCS Colloquia/AI

Speaker/Affiliation: Chitta
Baral/Arizona State University

Talk Audience: UTCS Faculty, Grads an

d Undergrads, Outside Interested Parties

Date/Time: Friday, March 25

, 2011, 11:00 a.m.

Location: ACES 2.402

Host: Vladimir Lifschit

z

Talk Title: Translating English to KR languages using inverse lambda
and parameter learning

Talk Abstract:
Our long term goal is to deve

lop general methodologies to translate natural language text into a formal

knowledge representation (KR) language. Our approach is inspired by Montagu

e’s path breaking thesis (1970) of viewing English as a formal language a

nd the research in natural language semantics. Our approach is based on PCC

G (Probabilistic Combinatorial Categorial Grammars), λ-calculus and stati

stical learning of parameters. In an initial work, we start with an initia

l vocabulary consisting of λ-calculus representations of a small set of wo

rds and a training corpus of sentences and their representation in a KR lan

guage. We develop a learning based system that learns the λ-calculus repre

sentation of words from this corpus and generalizes it to words of the same
category. The key and novel aspect in this learning is the development of

Inverse Lambda algorithms which when given λ-expressions β and γ can com

e up with an α such that application of α to β (or β to α) will give u

s γ. We augment this with learning of weights associated with multiple mea

nings of words. Our current system produces improved results on standard co

rpora on natural language interfaces for robot command and control and data

base queries. In an ongoing work we are able to use patterns to make guesse

s regarding the initial vocabulary. This together with learning of paramete

rs allow us to develop a fully automated (without any initial vocabulary) w

ay to translate English to designated KR languages. Our overall system is a
good example of integration of results from multiple sub-fields of AI and

computer science: machine learning, knowledge representation, natural lan

guage processing, λ-calculus (functional programming) and ontologies.

Speaker Bio:
Chitta Baral is a professor at the Arizona State Universi

ty. He obtained his B.Tech(Hons) degree from the Indian Institute of Techno

logy, Kharagpur in 1987 and his M.S and Ph.D degrees from the University o

f Maryland at College Park in 1990 and 1991 respectively. Chitta''s researc

h interests are in the areas of Artificial Intelligence, Knowledge Represe

ntation, Cognitive Robotics, Logic Programming, Natural Language process

ing and application of all that to Molecular Biology. His research has been
supported over the years by National Science Foundation, NASA, Science F

oundation Arizona, United Space Alliance, ONR, and ARDA/DTO/IARPA. He re

ceived the NSF CAREER award in 1995. He authored the book ``Knowledge Repre

sentation, Reasoning, and Declarative Problem Solving'''' published by Ca

mbridge University Press. He was an associate editor of the Journal of AI R

esearch and is an area editor of the ACM Transactions on Computational Logi

c. His recent research focus is on temporal specification of goals, reason

ing about actions and change in the multi-agent domain, combining probabil

istic and logical representation and reasoning, and most recently, on nat

ural language understanding through a learning based approach of translatin

g natural language to knowledge representation languages.