Professor Leonard M. Uhr
Professor Uhr earned his Ph.D. from the University of Michigan in
1957, served several years on the faculty there, and moved to the
University of Wisconsin faculty in 1965. As one of the initial
members of the UW Computer Sciences Department (which had
been established only a year prior to his joining), Professor Uhr was
instrumental in initiating the Department's curriculum and research
in artificial intelligence, and he was centrally influential in shaping
their maturation and evolution over his entire 26 years as an active
faculty member. He regularly taught both undergraduate and
graduate courses, was a successful research scientist, and was a
much-sought-after mentor of graduate students. He retired from
teaching in 1992 but remained active in research and publishing.
Professor Uhr's research and writing focused on computer
perception and learning, and on the use of parallel computer
architectures for artificial intelligence in general and for computer
vision in particular. He was quite expert in many aspects of human
neurophysiology and perception, and a central theme of much of his
research was to design computational structures and processes
based on his understanding of how the human nervous system
works. He was one of the early proponents of integration of symbolic
artificial intelligence with methods for dealing with uncertainty.
On these topics, Professor Uhr published seven books (as author
and/or editor) and nearly 150 journal and conference papers. His
seminal work was perhaps an article written in 1963 with Charles
Vossler, "A Pattern Recognition Program That Generates,
Evaluates, and Adjusts Its Own Operators," reprinted in
Computers and Thought --
edited by E. Feigenbaum and J. Feldman
-- which showcases the work of the scientists who
not only defined the field of artificial intelligence but who are responsible for having developed it into what it is today.
as Ph.D. major professor for 20 students, many of whom have gone
on to become in their own right important contributors to artificial
intelligence research and related areas of computer science.
Professor Uhr was a true intellectual, an independent thinker, a
scholar in love with ideas new and old, and a gentle man. His
passion for developing new insights made knowledgeable use of
and always manifested genuine respect for the thoughts and
perspectives of others -- whether those of his students, his UW
colleagues, or his colleagues and intellectual predecessors
Professor Uhr was born in 1927, and as a child, he attended the Oak Lane
Country Day School outside of Philadelphia. He graduated from Princeton in
1949, with a B.A. in Psychology. He received masters degrees in Philosophy
from the University of Brussels and Johns Hopkins University in 1951 before
obtaining his Ph.D. in Psychology from the University of Michigan in 1957.
Leonard Uhr was elected as a Fellow of the American Association for Artificial
Intelligence in 1995 for his pioneering contributions to artificial intelligence, but out of modesty, declined the honor.
Leonard Uhr died on
October 5, 2000 at his home in Madison. He was 73 years old. He is sorely missed
by those who had the good fortune to have known him.
What follows is a brief record of Leonard Uhr's professional life:
Ph.D., Psychology, University of Michigan, 1957.
M.A., Philosophy, Johns Hopkins Univ., 1951.
B.A., Psychology, Princeton Univ., 1949.
Professor, Computer Sciences Department, 1965-1992.
University of Wisconsin-Madison. Founded and led the Artificial Intelligence
(AI) group and the AI lab.
Several graduates of the program have made major contributions to AI.
significant research contributions to AI ranging from ground-breaking and
visionary early work in pattern recognition, learning, intelligent agent
modeling and parallel production systems to major contributions to parallel
architectures and languages for AI and vision, connectionist models,
brain modeling, and integration of symbolic and connectionist AI.
establish the interdepartmental Neuroscience program and was a member of
the Neuroscience faculty (1972-83)
- Director of AI lab, 1982-86
of Image Processing Lab, 1980-83.
1977-78, Honorary Senior Research Fellow, University College, London.
1963-65, Associate Professor, Psychology, University of Michigan.
1961-65, Chairman of Psychological Sciences, Mental Health Research Institute,
University of Michigan.
1960-65, Consultant, Systems Development Corporation and RAND Corporation.
1951-61, Several teaching and research positions.
Fullbright scholarship (1949-50)
Noteworthy Research Articles (out of a total of over 150)
- Honavar, V. & Uhr, L. (Ed.) Artificial Intelligence and Neural Networks:
Steps Toward Principled Integration. New York: Academic Press. 1994.
Uhr, L. Multi-Computer Architectures for Artificial Intelligence: Toward
Fast, Robust, Parallel Systems. New York: Wiley. 1987.
Uhr, L. (Ed.) Parallel Computer Vision. Boston: Academic Press. 1987.
Uhr, L. Algorithmically Structured Computer Arrays and Networks:
Architectures for Images, Percepts, Models, Information. Boston:
Academic Press. 1984.
Uhr, L. Pattern Recognition, Learning, and Thought. Englewood Cliffs:
Uhr, L. Pattern Recognition. (Ed.) New York: Wiley. 1966.
- Honavar, V. & Uhr, L. Integrating Symbol Processing and Connectionist
Networks, and Beyond. Invited Paper In: Intelligent Hybrid Systems.
Goonatilake, S. & Khebbal, S. (Ed.) London, Wiley: 1995.
Uhr, L. Digital and Analog Subnet Structures for Connectionist Networks.
In: Artificial Intelligence and Neural Networks: Steps Toward Principled Integration. Honavar, V. & Uhr, L. (ed.) Boston: Academic Press: 1994.
Honavar, V. & Uhr, L. Generative Learning Structures and Processes for
Connectionist Networks. Information Sciences 70 75-108. 1993.
Uhr, L. Forms Structure Form at Ever ``Higher'' and ``Lower'' Levels. In: Arcelli, C., Cordella, L.P., Sanniti di Baja, G. (ed). Visual Form: Analysis and Recognition. New York: Plenum. 1991.
Honavar, V. & Uhr, L. Coordination and Control Structures and Processes:
Possibilities for Connectionist Networks. Journal of Experimental and
Theoretical Artificial Intelligence 2 277-302. 1990.
Uhr, L. Increasing the Power of Connectionist Networks, By Improving
Structure, Processes, and Learning. Connection Science 2 179-193. 1990.
Li, Z. N. & Uhr, L. Pyramid Vision Using Key Features to Integrate
Image-Driven Bottom-Up and Model-Driven Top-Down Processes. IEEE Trans.
Systems, Man, and Cybernetics 16 250-262. 1987.
Li, Z.N. & Uhr, L. Evidential Reasoning in a Computer Vision System. In: Proceedings of the Second Annual Conference on Uncertainty in Artificial Intelligence. 1986.
Uhr, L. Massively Parallel Hardware-Software Structures for Learning.
In: Complex Systems - Operational Approaches to Neurobiology, Physics,
and Computers. H. Haken (Ed.), pp. 212-224. Berlin: Springer-Verlag. 1985.
Uhr, L. & Schmitt, L. The Several Steps from Icon to Symbol Using
Structured Cone/Pyramids. In: Multi-Resolution Systems for Image
Processing. A. Rosenfeld (Ed.), pp. 86-100. Amsterdam: North-Holland. 1984.
Uhr, L. Comparing Serial Computers, Arrays, and Networks Using Measures of
``Active Resources''. IEEE Transactions on Computers 31(10): 1022-1025. 1982
Uhr, L. & Douglass, R. A Parallel-Serial Recognition Cone System for Perception: Some Test Results. Pattern Recogition 11(1): 29-39. 1979.
Uhr, L. Parallel-Serial Production Systems With Many Working Memories. In:
Proc. IJCAI., Japan, 1979.
Uhr, L. & Kochen, M. Toward a Greater Generality in Artificial Intelligence. In: Proc. ECAI, pp. 351-354. Hamburg, Germany, 1978.
Uhr, L. & Kochen, M. Toward Adaptive (Computer-Based) Hospital Care Systems
That Can Grow and Improve as a Result of User Participation. International
Journal of Biomedical Computing, 10 191-203, 1979.
Uhr, L. Tryouts Toward the Production of Thought. In: Perception & Cognition: Issues in the Foundations of Psychology. Minnesota Studies in the Philosophy of Science. C.W. Savage (Ed). pp. 327-364. 1978.
Uhr, L. Toward Integrated Cognitive Systems, Which Must Make Fuzzy
Decisions About Fuzzy Problems. In: Fuzzy Sets. L. Zadeh et al. (Ed.),
pp. 353-393. New York: Academic Press, 1975.
Uhr, L. DECIDER-1, A System That Chooses Among Different Types of Acts.
In: Proc. 3rd IJCAI, Palo Alto, 1973.
Uhr, L. Flexible Syntactic Pattern Recognition. Pattern Recognition
3, 363-384, 1971.
Uhr, L. Layered Recognition Cone Networks That Preprocess, Classify, and
Describe. IEEE Trans. on Computers, 21, 758-768, 1972.
Uhr, L. & Jordan, M. The Learning of Parameters for Generating Compound
Characterizers for Pattern Recognition. In: Proc. IJCAI, Washington, 1969.
Uhr, L. & Kochen, M. MIKROKOSMS and Robots. In: Proc. IJCAI, Washington,
Uhr, L. Pattern-String Learning Programs, Behavioural Science 9, 258-270.
Uhr, L. Pattern Recognition Computers as Models for Perception, Psychogical
Bulletin 60 40-73, 1963.
Uhr, L. & Vossler, C. A Pattern Recognition Program That Generates,
Evaluates, and Adjusts its Own Operators. In: Computers and Thought,
E. Feigenbaum & J. Feldman (Ed.), New York: McGraw-Hill. 1963.
Selected Professional Activities
Editorial Board Member, Journal of Cognitive Systems Research (1999-2000)
Editorial Board Member, Pattern Recognition (1969-1993)
Editorial Board Member, Human Information Systems Management (1981-1987).
Founder and Co-Editor-in-Chief, Journal of Parallel and Distributed Computing
Member and Consultant, NSF Advisory Panel on University Computing Facilities
Member of numerous NSF Panels, Conference and Workshop Program Committees, etc.
Leonard Uhr's Research Contributions
Len Uhr was one of the earliest researchers in the field of Artificial
Intelligence. His contributions to the field span over 3 decades, on
topics as diverse as pattern recognition, machine learning, perception-
mediated reasoning and action, computer vision, neural networks, and
parallel and distributed architectures for AI.
Len was among the pioneers who worked on synthesis of computer programs
for tasks where neither the shape (form) nor the size (complexity) of
the solution can be specified a-priori. His paper with Vossler on
programs that generate, evaluate and adjust their own operators for
pattern recognition (published in 1963 in the volume "Computers and Thought",
edited by Feigenbaum and Feldman) is an example of such work that anticipated
much current research on automated program synthesis, constructive induction,
constructive neural network
Len's work with Kochen on simulation of agents published in the 1969 IJCAI
foreshadowed current research in Intelligent Agents.
He was among the first to propose parallel production system models for
perception, inference, and action in his 1971 IJCAI paper. This was followed
by more than a decade of significant research on various aspects of massively
parallel computer architectures and algorithms (including brain-like networks
of simple processors or connectionist networks - long before they became
fashionable) for computer vision, and perception-mediated reasoning and
Around 1986, Len's attention started to shift back to learning. Between 1986
and 1992, Len and his students developed some of the first approaches to
generative or constructive learning algorithms for pattern recognition aimed
to address some of the limitations of connectionist learning models
that relied entirely on parameter modification.
Between 1992 and 1996, Len occupied himself with several foundational problems in AI, including
the design of general, robust, and flexible adaptive architectures
for embodied intelligence and integration of symbolic and connectionist
approaches to AI and cognitive modelling.
During the past few years, Len focused his attention to societal and
educational impact of information technologies such as the Internet and the
World-Wide Web. He became very interested in developing tools for improving
access to information, bridging the gap between information haves and have nots
in the digital age, using information technologies to improve democracy, and
related societal issues. Most of his thoughts on this topic are
available only in the form of the notes that he kept and recollections of those
who had conversations with him. It is hoped that these records will be
made available to a larger audience in a suitable form in the near future.
Len founded the AI program at University of Wisconsin-Madison.
The UW-Madison AI group has produced a number of Ph.D.s who have gone on
to make significant contributions in artificial intelligence, computer science,
and even the arts, in academia as well as the industry.
As a professor, Len has, over the years, nurtured, fostered, and
mentored a large and diverse reservoir of talent in AI. The depth and breadth
of his knowledge in not only AI but also the related fields of psychology,
philosophy, and neuroscience, and his single-minded dedication to scientific
research and teaching has been truly inspiring to his many graduate students.
Ph.D. Dissertations Completed Under Leonard Uhr's Supervision
Timmreck, Eric M.
Advising by Computers: Course Advising, Medical Treatment, General Advising
July 3, 1968
Fabens, William J. H.
An Adaptive Interactive Teaching System for Programming Languages
August 13, 1970
Several Methods of Concept-Formation by Computer
December 12, 1969
Wexler, Jonathan D.
A Generative, Remedial and Query System for Teaching by Computer
March 21, 1970
Zobrist, Albert L.
Extraction and Representation of Features for Pattern Recognition and the Game of GO
August 10, 1970
Naylor, William Clark
Some Studies in Interactive Machine Imitation Using Character Recognition
February 23, 1971
Jordan, Sara R.
Learning to Use Contextual Patterns in Language Processing
September 7, 1971
Krueger, Myron William
Computer Controlled Responsive Environments
July 22, 1974
Williams, Harold A.
A Net-Structure Learning System for Pattern Description
December 12, 1973
August 7, 1974
Motion Extraction and Utilization in Scene Description
September 18, 1975
Douglass, Robert J.
A Computer Vision Model for Recognition, Description, and
Depth Perception in Outdoor Scenes
January 6, 1978
Korn, Robert K.
Machine Learning During Continuous Interaction With A
December 19, 1977
Orgren, Paul J.
The Induction of the Syntax of Natural Language by Computer
December 7, 1979
Schmitt, Lorenz A.
A Structured Approach to Computer Image Understanding; The Use and
Representation of Real-World Knowledge in an Artificial Vision System
June 21, 1982
Pyramid Vision Using Key Features and Evidential Reasoning
July 14, 1986
Learning Object-Centered Representations
July 30, 1987
Representing and Using Functional Definitions for Visual Recognition
December 10, 1987
Generative Learning Structures for Generalized Connectionist Networks
August 14, 1990
Learning Language about Objects and Using This Language to
Learn Further: The Childlike System
August 9, 1993
To be completed.
Professor Leonard Uhr's Obituary from the Wisconsin State Journal
This page is maintained in memory of Professor Leonard Uhr by:
Dr. Vasant Honavar
Director, Artificial Intelligence Research Laboratory
Department of Computer Science
Iowa State University
Atanasoff Hall, Ames, IA 50011-1040 USA
phone: +1-515-294-4377, fax: +1-515-294-0258