1. Daniel Anthony Oblinger
IBM T.J. Watson Research Citizenship: United States
19 Skyline Drive, Hawthorne, NY 10532 Phone: (917) 494-1272
Web: http://pobox.com/~oblinger E-mail: oblinger@pobox.com
EDUCATION
Ph.D. in Computer Science at the University of Illinois at Urbana-Champaign
M.S. in Computer Science from Ohio State University, GPA 4.0
B.S. in Mathematics and Computer Science from Northern Kentucky University
Summa Cum Laude • GPA 3.92 overall • GPA 4.0 in both majors
Thesis: “Plausible Inference: A knowledge-intensive, inductive approach to domain modeling.”
Doctoral Committee: Gerald DeJong (chair), Shankar Subramaniam, Larry Rendell, and Caroline Hayes
RESEARCH INTERESTS
Areas: Machine Learning, Data Mining, Artificial Intelligence.
PROFESSIONAL EXPERIENCE
Programming By Demonstration. Research Staff Member. IBM TJ Watson Research. (2002 - present)
• Won 12 person-years of "Adventurous Research" IBM funding
(sought-after funding is the most independent long-range funding offered within IBM)
• Approach combines my bioinformatics background in sequence alignment with structure-based
learning algorithms to learn structured scripts with loops and conditionals.
Scripts are automatically learned by passively observing human operators in action.
• On going efforts have yielded a dozen patent filings, publications, and two spin off projects.
• One spin off is aimed at IBM product sales driven by automatically generated product walk throughs.
• Second spin off integrates personal wizards into the IBM-owned Lotus Rich Client platform as an
end-user customization tool.
Statistical Pattern Recognition. Adjunct Faculty. Columbia University. (2002-3)
• Taught graduate-level courses in Machine Learning.
• Covered underlying theory and practical application of fourteen state of the art learning techniques.
• Class project involved open-ended research based on IBM dataset.
Email Mining. Research Staff Member. IBM TJ Watson Research. (2001-2)
• Project lead for two email mining systems: Mail Assistant facilitated personal contacts visualization
and retrieval based on email content and from/to graph linkages.
• Skills Miner utilized skill-related documents to learn a skill-word dictionary. This in turn drove
multi-instance learning approach for assessing employee skills based on email traffic patterns and
content.
• Side Activity: Co-created an IBM Research worldwide AI special interest group.
Speech Data Mining. Research Staff Member. IBM TJ Watson Research. (1998-2000)
• Developed speech analysis tool enabling teachers to identify reading progress and reading problems
in beginning readers. Tool relied on data from a companion system that read for and with children.
• In daily use in bi-lingual education programs at over 100 schools across the United States.
Protein Sequence Modeling. Research Assistant with S. Subramaniam. University of Illinois. (1995-7)
• Data mining consultant for a group of computational biologists.
• Learned protein similarity metrics based on atom-level descriptions of protein fold classes.
• Encoded approximate physical & chemical models of atomic interaction in order to constrain learning.
• Implemented an inductive learning algorithm that used these models to mine rules from 4 Gigs. of data
on protein structures. Implemented parallel version of algorithm for a 512-node CM5 supercomputer.
Learning/Inferencing Algorithms. Research Assistant with Gerald DeJong. Univ. of Illinois. (1989-95)
2. • Developed formal and computational models of plausible inference. This technique combines
a general but inaccurate expert model with specific and accurate entries from a database to learn
accurate and general rules describing the data.
• System Administrator for the AI group's file/mail/web/printer/xdm services.
Designed, scripted, and maintained a three-level backup scheme for this 50-seat network.
Software Engineer. IBM T.J. Watson Research Center. (1989)
• Collaborated on the initial design of an expert system that automatically generates configurations for
IBM mainframe computers.
Teaching Associate. Ohio State University. (1987-9)
• Designed and taught a course new to OSU’s curriculum: LISP for Engineers (CIS 459).
• Instructor: Pascal for Engineers (CIS 221; four courses).
• Invited to teach an OSU-accredited, off-campus version of Pascal (CIS 221) at Bell Core Corp.
Selected by department chair from over 60 candidates based on my previous student evaluation scores.
Research Associate. Ohio State University. (1988)
• Developed the semantics, designed, and implemented a data-directed control flow mechanism used to
drive part of the AI Tool Set, upon which many AI applications have been written.
Runtime Library Developer. IBM Santa Theresa Lab. (1988)
• Designed and implemented task synchronization primitives used in a run-time library that supports
more than five languages on the IBM 370.
Physical Attendant, Tutor, Engineer. Doug Ragland. Lexington, KY. (1984-6)
• Attendant for a mute quadriplegic. Assisted in all functions, including dressing, eating, swimming,
transportation, taking class notes, etc. One on one tutored for four computer science courses.
• Designed and implemented a text processing and programming environment for quadriplegics.
Environment included an extensible programmable editor, hierarchical note retrieval facility, and an
auto-word-completion dictionary; system designed to minimize keystrokes.
PROGRAMMING SKILLS
Java, C/C++, LISP/CLOS, Perl, XML, programming interfaces (Windows, X-windows, UNIX), a number
of more specific languages and many assemblers.
COLLEGIATE HONORS, ACTIVITIES, AND AWARDS
• Cognitive Science / Artificial Intelligence Fellowship (UIUC 1994)
• ACM Regional Programming Contest in Kalamazoo (1986) and Pittsburgh (1988)
Participated on OSU’s team (6th place out of over 50 schools) and NKU’s team (7th place)
• Student of the year in both mathematics and computer science at NKU (1987)
• NKU Dean’s Scholarship junior and senior years (1985–87)
• Placed in the top third in the annual Putnam Mathematics Competition (NKU 1985)
• AHP Mathematics Competition: second place as freshman (1984) and first place as sophomore (1984)
• NKU Mathematics Departmental Scholarship (1984–87)
• Kentucky State Science Fair: second place in physics and fourth overall (1982)
• Volunteer at the McKinley Foundation’s Emergency Men’s Shelter of Champaign (1995–96)
• President of the NKU Computer Science Club (1987)
• Captain for a team in the NKU intramural volleyball league (1986)
3. Daniel A. Oblinger
PROFESSIONAL ACTIVITIES
WORKSHOP/CONFERENCE CHAIR
“Workshop on human-understandable machine learning.” Twentieth National Conference on Artificial
Intelligence. (will be held July 9, 2005.)
“What works well where workshop” The Seventeenth International Conference on Machine Learning.
Stanford, CA. July 2000.
“Inductive Learning track” International Conference on Artificial Intelligence 2000.
Las Vegas, NV. June 2000.
“The Joint Beckman Institute / Hitachi Advanced Research Laboratory’s Symposium on Artificial
Intelligence” workshop at The Beckman Institute. Champaign, IL. May 1997.
ASSOCIATE EDITOR
International Conference on Artificial Intelligence. 2000.
ADJUNCT FACULTY
Co Professor for “ELEN E 6880 Statistical Pattern Recognition” at Columbia University, 2002-3.
REVIEWER
Machine Learning Journal special issue on Meta Learning. 2004.
The Fourth IEEE International Conference on Data Mining. 2004.
International Conference on Machine Learning. 2003.
The Third IEEE International Conference on Data Mining. 2003.
The Second IEEE International Conference on Data Mining. 2002.
The International Conference on Artificial Intelligence. 2000.
International Conference on Artificial Intelligence in Education. 2001.
International Joint Conference on Artificial Intelligence. 2001.
European Conference on Artificial Intelligence. 1994.
National Conference on Artificial Intelligence Student Program. 1993.
AFFILIATIONS
Advisory Board Member. Electrical and Computer Science Department, Northwestern University.
AAAI. American Association for Artificial Intelligence. 1991–
Phi Beta Kappa Honor Society. 1995–
ACM. Association for Computing Machinery. 1986–89
4. Daniel A. Oblinger
ISSUED PATENTS
6,873,990 Customer self service subsystem for context cluster discovery and validation.
6,853,998 Subsystem for classifying user contexts.
6,785,676 Subsystem for response set ordering and annotation.
6,778,193 Iconic interface for portal entry and search specification.
6,701,311 System for resource search and selection.
6,693,651 Iconic interface for resource search results display and selection.
6,643,639 Subsystem for adaptive indexing of resource solutions and resource lookup.
PUBLICATIONS
JOURNAL ARTICLES
R. Vilalta, D. Oblinger, "Evaluation metrics in classification: A quantification of distance-bias"
Computational Intelligence, Vol. 54, No. 3, pp. 187-193. 2003.
D. Oblinger, M. Reid, M. Brodie, R. Braz, "Cross Training and its application to skill mining"
IBM System Journal, Vol. 41, No. 3 pp. 449-460. 2002.
FIRST TIER CONFERENCES
T. Lau, L. Bergman, V. Castelli, D. Oblinger, “Sheepdog: Learning Procedures for Technical Support,”
Proceedings of the 2004 International Conference on Intelligent User Interfaces (IUI 2004). Madeira,
Portugal. 2004. pp. 109-116.
N. Mishra, D. Oblinger, and L. Pitt, “Sublinear time approximate clustering”
Proceedings of the Twelfth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA 2001).
San Francisco, CA. 2001.
R. Vilalta, D. Oblinger, “A Quantification of Distance-Bias Between Evaluation Metrics In Classification.”
Proceedings of the Seventeenth International Conference on Machine Learning, Stanford, CA. 2000. pp.
1087-1094.
D. Oblinger, G. DeJong, “An Alternative to Deduction.” Proceedings of the Thirteenth Annual Conference
of the Cognitive Science Society. Chicago, IL. 1991. pp. 837–841.
K. Forbus, D. Oblinger, “Making SME Greedy and Pragmatic.” Proceedings of the Twelfth Annual
Conference of the Cognitive Science Society. Cambridge, MA. 1990. pp. 61–68.
BOOK CHAPTER
5. Daniel A. Oblinger
G. DeJong, D. Oblinger, “A First Theory of Plausible Inference and Its Use in Continuous Domain
Planning.” Machine Learning Methods for Planning, Steven Minton (ed.). San Mateo, CA: Morgan
Kaufmann. 1993. pp. 93–124. (Selected from the Symposium on Learning Methods for Planning and
Scheduling for publication in extended book form.)
OTHER CONFERENCES, WORKSHOPS, AND SYMPOSIA
T. Lau, L. Bergman, V. Castelli, D. Oblinger, “ Programming Shell Scripts By Demonstration,''
The Nineteenth National Conference on Artificial Intelligence (AAAI 2004): Supervisory Control of
Learning and Adaptive Systems workshop. San Jose, CA. 2004.
L. Bergman, T. Lau, V. Castelli, D. Oblinger, “Programming-by-demonstration for Behavior-based User
Interface Customization,” Proceedings of the Workshop on Behavior-Based User Interface
Customization, (IUI 2004). Madeira, Portugal. 2004.
T. Lau, D. Oblinger, L. Bergman, V. Castelli, C. Anderson, “Learning Procedures for Autonomic
Computing,” Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence
(JCAI 2003): Developing a Research Agenda for Self-Managing Computer Systems workshop,.
Acapulco, Mexico. 2003.
L. Bergman, T. Lau, V. Castelli, D. Oblinger, “Personal Wizards: Collaborative End-User
Programming,” Proceedings of the CHI 2003 Conference on Human Factors: Workshop on Perspectives
in End User Development, Fort Lauderdale, Florida. 2003.
R. Vilalta, M. Brodie, D. Oblinger, and I. Rish, "A Unified Framework For Evaluation Metrics In
Classification Using Decision Trees." Proceedings of the 12th European Conference on Machine
Learning (ECML 2001), Freiburg, Germany. 2001.
D. Gordin, R. Farrell, D. Oblinger, Mapping knowledge production in network organizations.
IBM Academy of Technology Conference on Knowledge Management. Zurich, Switzerland. April 2001
D. Oblinger, G. DeJong, “Dynamic-Bias Induction.” American Association for Artificial Intelligence
Fall Symposium Series on Relevance (AAAI-94). New Orleans, LA. 1994. pp. 164–67
G. DeJong, D. Oblinger, “A First Theory of Plausible Inference and Its Use in Continuous Domain
Planning.’’ Symposium on Learning Methods for Planning and Scheduling. 1991
INVITED TALKS AND SEMINARS
“Learning System Management Procedure By Demonstration.” Computer Science Colloquium, New
York University. November 2003.
“Personal Wizards: A Programming by Demonstration Approach” Northwestern University, April 2003.
“Knowledge-Based Induction of Protein Tertiary Structure.” The Joint Beckman Institute / Hitachi
Advanced Research Laboratory’s Symposium on Artificial Intelligence. Beckman Institute. August
1995
“Minimum Description Length Principle.” The Machine Learning Seminar Series. Beckman Institute.
April 1994
TECHNICAL REPORTS AND WORKING PAPERS
6. Daniel A. Oblinger
D. Oblinger, V. Castelli, T. Lau, L. Bergman. "Similarity-Based Alignment and Generalization: A New
Paradigm for Programming by Demonstration."
IBM T.J. Watson Research Center: Technical Report: RC23140. 2004.
V. Castelli, D. Oblinger; L. Bergman; T. Lau. "Dynamic Model Selection in IOHMMs"
IBM T.J. Watson Research Center: Technical Report: RC23395. 2004.
T. Lau, L. Bergman, V. Castelli, D. Oblinger, “Programming Shell Scripts by Demonstration”
IBM T.J. Watson Research Center: Technical Report: RC23218. 2004.
L. Bergman, T. Lau, V. Castelli, D. Oblinger. "Programming-by-Demonstration for Behavior-based User
Interface Customization" IBM T.J. Watson Research Center: Technical Report: RC23116. 2004.
T. Lau, D. Oblinger, L. Bergman, V. Castelli, C. Anderson. "Learning Procedures for Autonomic
Computing"
IBM T.J. Watson Research Center: Technical Report: RC23115. 2004.
D. Oblinger, “Plausible Inference: A Knowledge-Intensive, Inductive Approach to Domain Modeling.”
University of Illinois: Technical Report UIUC-DCS-R-97-2004. Urbana, IL. 1997.
D. Oblinger, G. DeJong, “Towards an Inductive Model of Defeasible Inference.” Beckman Institute,
University of Illinois: Technical Report UIUC-AI-BI-95-01. Urbana, IL. 1995.
D. Oblinger, G. DeJong, “Dynamic Bias Induction.” Beckman Institute, University of Illinois: Technical
Report UIUC-AI-BI-9403. Urbana, IL. 1994. (Extended version)
D. Oblinger, G. DeJong, “An Alternative to Deduction.” Department of Computer Science, University of
Illinois: Technical Report UIUC-DCS-R-91-1688. Urbana, IL. 1991. (Extended version)
J. Josephson, D. Smetters, R. Fox, D. Oblinger, A. Welch, G. Northrup, “The Integrated Generic Task
Toolset—Fafter release 1.0—Introduction and User’s Guide.” The Ohio State University Laboratory for
Artificial Intelligence Research: Technical Report 89-JJ-FAFNER. Columbus, OH. 1992.
REFERENCES AVAILABLE UPON REQUEST