Problem Formulation for Computer-Supported Solving

Wednesday 11/12/2013 12:00-1:00PM
Location: SJ050
Speaker: Prof. Steve Tanimoto


The world is full of problems to be solved. However, most of them are
not posed in a form amenable to computer-based collaborative problem
solving. A problem must be formulated appropriately. This lecture
deconstructs the process of formulating a problem so that it can be
solved within a collaborative problem-solving system such as the
prototype system CoSolve, developed at the University of Washington.
The process will be illustrated with well-known problems in artificial
intelligence education, as well as some from the realm of design.
General approaches such as the use of state variables and constraints
will be described. A formulation of the posing process itself as a
problem leads to the notion of a meta-problem template.


Steven Tanimoto is a professor computer science and engineering and
adjunct professor of electrical engineering at the University of
Washington, Seattle. His past research focused on parallel
architectures and algorithms for image processing. Other research has
addressed aspects of visual languages that engage learners, and the
application of transparent interfaces to the design of software for
learning. He has served as the editor-in-chief of the IEEE
Transactions on Pattern Analysis and Machine Intelligence, helping to
build it into the leading journal in its field. As the PI of an NSF
project entitled Mathematics Experiences Through Image Processing, he
oversaw the development of software that included the original Pixel
Calculator application (intended for middle-school students), the
Color Pixel Calculator, and that led eventually to the development of
PixelMath2012, an integrated environment for exploring image
processing with formulas and Python programs (aimed at high-school and
college students). His recent text, An Interdisciplinary Introduction
to Image Processing: Pixels, Numbers, and Programs, (MIT Press, Spring
2012) is designed to engage students in the subject matter via its
connections to mathematics, computing, visual effects, and
applications. He recently directed the development of CoSolve, an
experimental online facility that mediates collaborative problem
solving. He is spending the Michaelmas term of 2013 as a visiting
researcher at the Computer Laboratory of Cambridge University.