|
|
|
|
|
| David E. Johnson (manager) |
  |
|
|
Education: PhD in Theoretical Linguistics,
University of Illinois, Urbana-Champaign,
1974.
I first came to the T.J. Watson Research
Center as a visiting scientist in 1974 to
pursue research in theoretical linguistics,
but earned my keep as a part-time grammar
writer for the IBM Transformational Question
Answering system. I wound up staying at Watson
until 1979, working mostly on relational
grammar and ergative languages. During 1980-82,
I taught linguistics at Yale University and
then worked on machine translation at Yale's
Artificial Intelligence Lab. In 1982, I came
back to Watson as a research staff member,
devoting most of my time to the development
of portable natural-language database query
systems. 1987-1989 was spent at the IBM Tokyo
Research Lab, where I worked officially on
the IBM JETS Japanese-English machine translation
system and unofficially on my Japanese. Returning
to Watson in 1989, I switched my attention
to a succession of topics including feature
structure logics, linguistically enabled
interactive multimodal user interfaces, and
most recently, machine-learning based text
categorization and information extraction
systems. These days I describe myself as
a "funny kind of engineer". However,
as my publications indicate, I have not completely abandoned
theoretical linguistics.
|
|
|
| Fred Damerau |
|
|
|
Education: Cornell U., BA Mathematics; Yale
U., M.A., Ph.D. Linguistics.
Joined IBM in 1957 to work on air defense
computers, then Intelligence Data Handling
system for SAC, USAF. Joined research in
1961. Worked on text processing (spelling
correction, hyphenation), information retrieval,
question answering, customization, machine
learning for text categorization, information
extraction and a variety of other problems.
|
|
|
| Thilo Goetz |
|
|
|
Education: PhD in Computational Linguistics,
U. of Tübingen, Germany.
Joined IBM Research in 1997. Current and
past interests include natural language syntax
and semantics; a wide spectrum of parsing
methods, from finite state to constraint
based; information extraction; logic and
complexity theory; and software technology
and standards for NLP.
|
|
|
| Sylvie Levesque |
|
|
|
Education: U. of Quebec in Montreal, Canada.
B.Sc, M.Sc. Computer Science.
Joined IBM Canada in 1996 as IT/Architect
after working in several R&D organizations.
On international assignment at Watson Research
from 1997 to 2001 in the Conversational Machines
group to work on various Natural Language
Dialog projects and then with the Computational
Linguistics and Text Mining group. Joined
Research in 2001. Contributed to the IBM
Web Guide effort ( aka WebGenie ) and worked
on later TextAnalyzer extensions. Current
work is on Interactive Learning Environments
and Information Extraction.
|
|
|
| Frank J. Oles |
|
|
|
Education: More than you can shake a stick
at, culminating in a Ph.D. in Computer and
Information Science from Syracuse University.
Oles is an IBM Research Staff Member, vintage
1983, specializing in finding actual uses
of work on the mathematical foundations of
computer science, particularly aspects related
to universal algebra, category theory, ordered
structures, general topology, and logic.
(N.B. You often have to do the foundational
work before you can use it. Only rarely is
it just sitting on the shelf waiting for
a use.) In the past, he worked on the mathematical
semantics of programming languages. Later,
he worked on the K-REP knowledge representation
language, defining its subsumption algorithm
based on the algebraic structure of the language.
In 1996, he migrated to the Computational
Linguistics and Text Mining Group, where
there was plenty of new stuff to work on,
including the prototype for the IBM Text
Analyzer, a decision-tree-based symbolic
rule induction system for text categorization.
He found that ordered structures could be
used in defining dialog management in the
IBM Web Guide, a system supporting natural
language interaction between a user and a
website. And, most recently, he has been
working on new ideas for pattern generalization
that should prove useful for relational learning,
in general, and for information extraction
from text, in particular. He did other stuff,
too, but we have to keep this short.
|
|
|
| Brian F. White |
|
|
|
Joined IBM in 1982, as a member of the TQA
natural language query group.
Developed a grammar-based SQL paraphraser,
and worked on pronouns and anaphora, and
on tokenization and morphology; late in the
project, was responsible for the analysis
grammar. On assignment to the Sylvia natural
language query project in Stockholm, designed
a paraphraser for a variant of first order
logic, and supervised analysis and paraphraser
grammar development.
Helped to develop the existing version of
the K-REP terminological knowledge representation
system. Worked on a lightweight document
matcher and on graph-based approaches to
document clustering while in the Data Abstraction
Research group.
In the Computational Linguistics and Text
Mining group, current work is on LPC, which
supports Horn clause reasoning and full first-order
logic. Current focus is on automated reasoning
and knowledge representation.
|
|
|
| Tong Zhang |
|
|
|
Education: BA in mathematics and computer
science, Cornell U, NY; PhD in computer science,
Stanford U, CA.
Joined IBM in 1998. Research interests include
numerical analysis, statistical machine learning
and applications. Worked on text categorization
and helped to develop machine learning algorithms
for IBM text analyzer. Also worked on general
methodologies for statistical machine learning
and their applications in natural language
processing and information extraction.
|
|
|
|
