Research interests and publications
I am interested in using technology to enhance and to understand
creative processes. In particular I enjoy
applying computer science to problems centered around natural
language modeling and understanding. Sample applications include text
prediction, compression, summarization, classification, and
generation; transmission error detection and recovery; information
retrieval and extraction; speech and handwriting recognition; and
machine translation.
My other academic interests include parallel programming, theoretical
computer science, and computer-based education. My advisor at
Carnegie Mellon was John
Lafferty.
Below are some recent papers and presentations I've contributed
to. Please contact me before redistributing any of them.
Published journal papers
Published conference papers
- D. Beeferman. The rhythm of lexical stress in prose.
In Proc. Assoc
Computational Linguistics '96, Santa Cruz, CA.
[ Postscript ]
- D. Beeferman, A. Berger, and J. Lafferty.
A model of lexical attraction and repulsion.
In
Proceedings of the ACL-EACL '97 Joint Conference, Madrid, Spain.
[
Abstract ] [ Postscript preprint ]
- D. Beeferman, A. Berger, and J. Lafferty.
Text segmentation using exponential models.
In Proc. Empirical Methods in Natural Language Processing 2 (AAAI) '97,
Providence, RI.
[
Abstract and postscript preprint ]
- D. Beeferman, A. Berger, J. Lafferty. A lightweight
punctuation annotation system for speech. ICASSP '98.
[ Postscript ]
[ Demo ]
- D. Beeferman, A. Berger, Agglomerative clustering of s earch
engine query log. KDD 2000.
[ PDF ]
Workshop papers
- D. Beeferman. "Lexical discovery with an enriched semantic network".
Describes the evolving FreeNet database system, and
Lexical FreeNet.
Proceedings of the Workshop on Applications of
WordNet in Natural Language Processing Systems, ACL/COLING 1998.
[
PDF ]
- S. Chen, D. Beeferman, R. Rosenfeld.
Evaluation Metrics for Language Models.
Appeared at the Broadcast News Transcription and Understanding Workshop,
February 1998.
[ Postscript ]
[
HTML ]
- D. Beeferman. "A probabilistic error metric for segmentation algorithms".
An abstract for a talk I gave at the Speech and Language Technology
workshop in Sheffield, England, June 1997.
[ Postscript ]
Unpublished projects, notes, and talks
- D. Beeferman. "QPD: Query by Pitch Dynamics. Indexing
Tonal Music by Content".
A course project write-up discussing a method of indexing
and clustering computer music files by content.
[
PDF ]
- D. Beeferman. "Count your blessings...faster!".
A lecture on combinatorics and representation
I gave for Mathematical Foundations of Computer
Science on 2/17/1998.
[
HTML ]
[
Powerpoint ]
[
PDF ]
- D. Beeferman. "Exponential Models for Natural Language".
A talk I gave locally about my group's ongoing work with
exponential language models.
[ Postscript ]
- D. Beeferman. "A note on the water jug problem in Prodigy".
A discussion of the redesign of a Prodigy domain involving infinite
types.
[ Postscript ]
- D. Beeferman. Link Grammar Prefix Measures for Spontaneous Speech
Recognition. Originally a term project for
a speech understanding course, this paper
discusses present and future work
toward using the link grammar formalism in conversational
speech recognition.
[ Postscript ]
Book chapter
- Huang, X., A. Acero, F. Alleva, D. Beeferman, M. Hwang, and
M. Mahajan. From CMU Sphinx-II to Microsoft Whisper: Making Speech
Recognition Usable. Appeared in Advanced Topics in Speech
Recognition, Kluwer Academic Publishers, 1995.
Discusses the evolution of the Microsoft
WHISPER speech recognizer.
[ Microsoft Word 6.0 ]
Go back to Doug's homepage