IEEE Computer Society and GBC/ACM
7:00 PM, Thursday, September 18, 2008
Broad Institute Auditorium (MIT building NE-30)
Evolvability and Robust Design
Gerald Jay Sussman
It is hard to build robust systems: systems that have acceptable
behavior over a larger class of situations than was anticipated by
their designers. The most robust systems are evolvable: they can be
easily adapted to new situations with only minor modification. How
can we design systems that are flexible in this way?
Observations of biological systems tell us a great deal about how to
make robust and evolvable systems. Techniques originally developed in
support of symbolic Artificial Intelligence can be viewed as ways of
enhancing robustness and evolvability in programs and other engineered
systems. By contrast, common practice of computer science actively
discourages the construction of robust systems.
Robust designs are built on an additive infrastructure: there are
exposed interfaces for attaching new functionality without serious
disruption of preexisting mechanisms. Indeed, the ability to
harmlessly duplicate a mechanism and then modify the copy to supply
useful new functionality is one of the principal ploys appearing in
natural evolution. What are the preconditions that support such
augmentation? Can we engineers arrange our systems to be extensible
in this way? Are there exploitable analogies between the techniques
that we have created to make extensible artifacts and the mechanisms
that we find in biological systems?
One powerful idea in biological systems is the use of space to
distribute function. In principle, each cell of a multicellular
organism is capable of living by itself. But a multicellular organism
is a community of cells that are spatially differentiated to build
particular structures and to perform particular functions. How is
this arranged? We now understand how this can this contribute to
reliability, but do we understand how it contributes to flexibility?
I will address these and related issues, and show, in terms of
explicit and concrete examples, how some of the insights we glean can
inform the way we do engineering in the age of information.
Gerald Jay Sussman is the Panasonic (formerly Matsushita) Professor of
Electrical Engineering at the Massachusetts Institute of Technology.
He received the S.B. and the Ph.D. degrees in mathematics from the
Massachusetts Institute of Technology in 1968 and 1973, respectively.
He has been involved in artificial intelligence research at
M.I.T. since 1964. His research has centered on understanding the
problem-solving strategies used by scientists and engineers, with the
goals of automating parts of the process and formalizing it to provide
more effective methods of science and engineering education. Sussman
has also worked in computer languages, in computer architecture and in
VLSI design.
Sussman is a coauthor (with Hal Abelson and Julie Sussman) of the widely used
introductory computer science textbook,
"Structure and Interpretation of Computer Programs," which has been
translated into French, German, Chinese, Polish, and Japanese. As a
result of this and other contributions to computer-science education,
Sussman received the ACM's Karl Karlstrom Outstanding Educator Award
in 1990, and the Amar G. Bose award for teaching in 1991.
Sussman's contributions to Artificial Intelligence include problem
solving by debugging almost-right plans, propagation of constraints
applied to electrical circuit analysis and synthesis, dependency-based
explanation and dependency-based backtracking, and various language
structures for expressing problem-solving strategies. Sussman and his
former student, Guy L. Steele Jr., invented the Scheme programming
language in 1975.
Sussman saw that Artificial Intelligence ideas can be applied to
computer-aided design. Sussman developed, with his graduate students,
sophisticated computer-aided design tools for VLSI. Steele made the
first Scheme chips in 1978. These ideas and the AI-based CAD
technology to support them were further developed in the Scheme chips
of 1979 and 1981. The technique and experience developed was then
used to design other special-purpose computers. Sussman was the
principal designer of the Digital Orrery, a machine designed to do
high-precision integrations for orbital-mechanics experiments. The
Orrery was designed and built by a few people in a few months, using
AI-based simulation and compilation tools.
Using the Digital Orrery, Sussman has worked with Jack Wisdom to
discover numerical evidence for chaotic motions in the outer planets.
The Digital Orrery is now retired at the Smithsonian Institution in
Washington DC. Sussman was also the lead designer of the Supercomputer
Toolkit, another multiprocessor computer optimized for evolving
systems of ordinary differential equations. The Supercomputer Toolkit
was used by Sussman and Wisdom to confirm and extend the discoveries
made with the Digital Orrery to include the entire planetary system.
Sussman has pioneered the use of computational descriptions to
communicate methodological ideas in teaching subjects in Electrical
Circuits and in Signals and Systems. Over the past decade Sussman and
Wisdom have developed a subject that uses computational techniques to
communicate a deeper understanding of advanced Classical Mechanics.
Computational algorithms are used to express the methods used in the
analysis of dynamical phenomena. Expressing the methods in a computer
language forces them to be unambiguous and computationally effective.
Students are expected to read our programs and to extend them and to
write new ones. The task of formulating a method as a
computer-executable program and debugging that program is a powerful
exercise in the learning process. Also, once formalized procedurally,
a mathematical idea becomes a tool that can be used directly to
compute results. Sussman and Wisdom, with Meinhard Mayer, have
produced a textbook, "Structure and Interpretation of Classical
Mechanics," to capture these ideas.
Sussman is a fellow of the Institute of Electrical and Electronics
Engineers (IEEE). He is a member of the National Academy of
Engineering (NAE), a fellow of the American Association for the
Advancement of Science (AAAS), a fellow of the American Association
for Artificial Intelligence (AAAI), a fellow of the Association for
Computing Machinery (ACM), a fellow of the American Academy of Arts
and Sciences, and a fellow of the New York Academy of Sciences (NYAS). He is
also a bonded locksmith, a life member of the American
Watchmakers-Clockmakers Institute (AWI), a member of the Massachusetts
Watchmakers-Clockmakers Association, a member of the Amateur Telescope
Makers of Boston (ATMOB), and a member of the American Radio Relay
League (ARRL)..
This
joint meeting of the Boston/Central New England Chapter of the IEEE
Computer Society and GBC/ACM will be held in the Broad Institute
Auditorium (MIT building NE-30). The Broad Institute is on Main St
between Vassar and Ames streets. You can see it on a map at this location. The auditorium is on the ground floor near the entrance.
For more information contact Peter Mager
(p.mager at computer.org)
Updated:
July 15, 2008.