Curriculum descant: the AI education repository

Abstract

Association for Artificial Intelligence’s (AAAI) Fall Symposium series, the symposium “Improving the Instruction of Introductory AI” was held. Barbara Grosz, then president of AAAI, concluded the symposium by accepting the following two action items for the association (see SIGART Bulletin 6, 2, April 1995): 1. Have available at the AAAI Web site a central repository of programs, tools, assignments, and papers , and 2. Organize tutorials to be given at AAAI’s National Conferences and the International Joint Conferences on Artificial Intelligence (IJCAI) on how to teach the more specialized portions of the introductory artificial intelligence (AI) course, such as vision, NLP, and robotics. Five years have passed since that symposium and it is almost time for AAAI to host another one. In this installment, I would like to discuss the AAAI educational repository mentioned in Action Item 1. The repository is a central distribution point that (1) focuses on Web educational resources and materials to be used in undergraduate AI pedagogy, (2) identifies and organizes such resources by topic, and (3) provides diverse resources applicable to various levels of student expertise and budgetary considerations. The repository includes information on current AI textbooks, links to syllabi, sample programming and written assignments, online tutorials, information on tools and environments that may be used in the classroom, and papers on AI pedagogy. Also included are mechanisms for submitting personal AI education resources. The repository can be accessed at http://www.aaai.org by clicking on the link AI Resources and, under the heading “Source Information,” clicking on the link AAAI Educational Repository. The repository is currently maintained by Bill Manaris of the University of Southwestern Louisiana. At the aforementioned symposium, Pat Hayes and Ken Ford issued a challenge for the creation of a Web-based “textbook.” The chapters in the textbook would cover specific topics, each an up-to-date survey of a subarea of AI, written by people with thorough knowledge of that area. The idea was to have such a text continually revised and expanded to adapt to changes in AI. This multiauthor text would exist as distributed content managed by sufficient editorial control and commentary to ensure common vocabulary and style. Although I am uncertain about the progress on the challenge, I would like to have you consider using the educational repository as a possible infrastructure for the creation of such a text. Use it to obtain materials for which you may not have expertise, and contribute educational materials for the topics in which you do have expertise. As for organizing tutorials on teaching of specific topics (Action Item 2), I think AAAI and IJCAI could do a better job. Although some tutorial-like sessions have been held at these conferences, most focus on informing AI researchers and practitioners rather than educators. I believe that AAAI, and its membership, ought to extend beyond AAAI, IJCAI, and other topic-specific AI conferences to offer tutorials. Teaching of AI is largely carried out by computer science faculty who may not be directly involved in doing AI research. Possibly the best place to hold tutorials is at the annual meetings of ACM’s Special Interest Group on Computer Science Education (SIGCSE). SIGCSE meetings have a rich tradition of offering educators workshops on the teaching of new methodologies, technologies, and ideas emerging in computer science. More than half the attendees at the SIGCSE annual meetings enroll in these workshops also. For more information on ACM SIGCSE and its Workshop Program for educators, go to the SIG’s Web page (http://www.acm.org/sigcse). Deepak Kumar Department of Mathematics and Computer Science Bryn Mawr College Bryn Mawr, PA 19010 dkumar@brynmawr.edu