HM725*. Enactive Artificial Intelligence |
|
| Instructor | Kevin McGee |
| Semester | Spring 2004. Info from IDA |
| Prerequisites | Programming experience or background in cognitive science.
Graduate and D-level students in computer science or applied cognitive science. |
| Goals | To explore, understand, and model key issues of “enactive cognitive AI”. |
| Content | This is a project-oriented course for cognitive scientists
interested in modeling cognition in terms of “enactive” mechanisms.
Enactive AI is an outgrowth of "constructivist AI" and has its
roots in developmental psychology, constructivism, and self-organizing models
of biology. This approach to AI differs from others in that the emphasis
is not on the “recovery” of (pre-given) features of the world,
but rather how autonomous systems can generate viable “life-worlds”
through their activity. Students will be asked to form teams and develop
working implementations of key ideas from Enactive AI. The course will explore relevant concepts and mechanisms from Enactive AI, including: development psychology; autopoiesis; self-organization and self-maintenance; cellular automata; "situated action"; and subsumption architectures. |
| Literature | Readings will be short and distributed as needed, and may include work by Jean Piaget, Jakob von Uexkull, Marvin Minsky, Seymour Papert, Humberto Maturana, Francisco Varela, Eleanor Rosch, Susan Oyama, Gary Drescher, Barry McMullin, and Rodney Brooks. Note that "readings" will also include using/studying existing simulations. |
| Examination | Active participation, weekly deliverables, and a public presentation of a completed final project. |
| Other |
The course is organized as a series of design sessions, discussions, and small, weekly “deliverables.” The course will meet once a week for 8 weeks; each week will be divided between design sessions – and “status” presentations/discussions of student projects. The course will also meet at the end of the quarter for final presentations of student projects. Course size is limited to 20 participants. Course language is English.
Note: for this course we will try to create a balance between students
who want to build their own simulations -- and students who want to write
papers about the course themes. In both cases, the emphasis in this course
is on “completing a final project,” so students are expected
to be comfortable with either a) programming, or b) writing a publication-quality
paper. Students who apply for the course should send an email indicating
which kind of activity/project they would like to do. (For implementations,
students are free to choose any programming language -- Java, Scheme,
Lisp, etc. -- but a final project must work "for real" in some
significant sense. It cannot be a "mock up" or "Director
presentation.") |
| *) HMI-number may change | |