In the summer of 1993, for the first time, I taught the course, Complex
Intelligent Systems (CIS), based COMPLETELY on my published and yet unpublished
papers on Linguistic Geometry. I developed a 600-page set of slides, handouts,
and assignments. The handouts included some of my research papers. Since then, I
have taught this course 4 times. I was always very enthusiastic preparing and
teaching this course: this is the level of teaching I had in Russia, right at the
frontier of the unknown. Every time when I offer this course I completely
redevelop all the course materials. This may sound like a paradox, because this
is the subject I know best. Probably, just because of that, I am constantly
fighting with myself for better understanding and presenting this novel subject
to the students. I should not say "I am"; the truth is that my students are
fighting by my side.
From the very first offer, this course was a success. Even the students who never
took my introductory classes in Artificial Intelligence catch up so quickly that
they soon become the leaders in a race to solve problems on the board. Students
are usually excited about the topic and contents of this course. In fact, they
write about it in their unsolicited letters. During the lectures on CIS students
often suggest some corrections to the formal grammars, search trees, and
algorithms considered in this course, and I incorporate these corrections in my
papers and into the following versions of course materials. While preparing and
teaching this course I am taking another fresh look at the concepts of LG,
generate new ideas and they immediately find their way into the new research
papers. Basically, the experience of teaching this course has been constantly
used in my numerous extended presentations, tutorials, and short courses around
the world. Quite often when I teach a full or half day tutorial on LG at the
international conferences I present some concepts in a manner developed a week
before during my discussions with students in the CIS class.
In teaching CIS, I tried to show the students a variety of unexplored problems,
and a number of students expressed their intent to prepare a thesis on LG. While,
a number of them finished their work and defended their theses in the past years,
here I would like to distinguish four of them.
- Chris Fletcher, who is a manager at Lockheed Martin Astronautics, developed a
Software Simulator of robot control in the industrial environment. After the
highly successful defense at UCD, this Simulator was demonstrated at Sandia
National Laboratories, Rockwell International Corp., at the National Forum on LG
in San Marcos, TX, in 1997, and at the Second Symposium on LG in Berlin, Germany,
in 1997. The following development should generate an industrial prototype.
- Rick Turek, a former Lockheed Martin employee, who is currently with
Geographical Information Systems (GIS) Solutions, developed a System for real
time emergency vehicle routing. The prototype of this System was tested for fire
vehicles routing in Aurora, CO, with outstanding results, exceeding by far all
the systems that are currently in use. Various versions of the System were
demonstrated at the Goddard Conference on Applications of Artificial Intelligence
at Goddard Space Flight Center, MD, in 1995 and, also, presented at the National
Forum on LG in Texas in 1997. Moreover, GIS Solutions, the new company
established by Rick Turek and a group of other former Lockheed Martin employees,
will use this System as a prototype for developing commercial hardware/software
products for emergency vehicles routing.
- David Wood, a former employee of the Carnegie Software Group, developed a ³BGF:
A Framework for Board Game Applets² to be used for implementation of various
on-line LG tools on Java. The goal of this system is twofold: (1) exploration of
the new concepts of reusable software development on Java, and (2) creation of
the framework for an animated WEB site for on-line remote learning of LG concepts
via experiments with LG applications. BGF was demonstrated at Pioneer
Technologies and at Rockwell International Corp.
- The most interesting experimental LG system has been developed by Eduard
Skhisov. It is called War Game Simulation Tool. It is based on the newest papers
on LG and includes a pilot implementation of the no-search approach in LG for
concurrent combat of the aircraft. Following this approach the search for a
solution is replaced by the construction of winning strategies for each of the
opposing sides. This approach generates the optimal strategy, a problem solution.
A preliminary version of this software tool was demonstrated at the Second
Symposium on LG in Berlin, Germany and at Rockwell International Corp. in 1997. A
highly successful thesis defense at the University of Denver elicited interest of
many professors and students of two departments. In particular, this defense
included on-line demonstration (via WWW) of the combat simulation while the Tool
itself was located off-site on the remote network of computers (working in
parallel). It employed BGF Framework (on Java) developed earlier by David Wood. A
new advanced version of the Tool will be presented at the Third Symposium on LG
in San Diego, CA in October of 1998.
|
