SUMMARYProjects in the Artificial Intelligence course have evolved over the years. Along the way, they have taken several forms, including small-scale LISP/Prolog projects, larger-scale object-oriented projects in CLOS/C++, projects organized around games, and more recently, projects organized around the concept of agents. All along, educators have attempted to make the projects more appealing and instructive at the same time.In this panel, we will examine three disparate approaches for making AI projects more instructive and engaging: •The first approach organizes all the projects around a central theme, in this case, machine learning; •The second approach uses inexpensive robots as the platform for traditional projects; •The third approach moves to a software platform that enables working with advanced or simulated robots as well. All three approaches have been evaluated, and make supplementary materials available for use by interested faculty. MACHINE LEARNING -INGRID RUSSELLWe will present a suite of adaptable hands-on projects that can be closely integrated into a one term AI course. Our work unifies the Artificial Intelligence (AI) course around the theme of machine learning and creates an adaptable framework for presenting core AI concepts around that theme. Machine learning is inherently connected with the AI core topics and provides methodology and technology to enhance real-world applications within many of these topics. Machine learning is now considered as a technology for both software development (especially suitable for difficult-toprogram applications or for customizing software) and building intelligent software (i.e., a tool for AI programming). Our projects emphasize the relationship between AI and computer science in general, and software development in particular and highlight the bridge that machine learning provides between AI technology and modern software engineering. Each project involves the design and development of a learning system which will enhance a particular commonly-deployed application. In an introductory course one wishes to impart a wide variety of topics efficiently, indexing the major areas of the field. A machine learning application can be rapidly prototyped, allowing learning to be grounded in engaging experience without limiting the important breadth of an introductory course. The projects span several applications including Web user profiling, character recognition, the N-Puzzle problem, the jeopardy dice game Pig, Web document classification, and the popular board game Clue.We will present our projects as well as our experiences using them. Evaluation results indicate that the projects enhanced the student learning experience in the introductory AI course and that students demonstrated a better understanding of fundamental AI concepts such as Knowledge Representation and Search. Students were better motivated to learn the fundamental concepts both of AI and machine learning. The projects also stimulate students' interest in additional AI and machine learning related ...

The benefits of engaging undergraduate students in research are numerous and well-known. Therefore, many schools are encouraging undergraduate research. However, carrying out undergraduate research in a liberal arts school can be challenging -liberal arts schools usually lack the resources typically available in larger research universities; the research programs of faculty at such schools are often insular; undergraduates may not always be adequately prepared or motivated for research; and research is only one of the many activities competing for the time and energy of undergraduate students. The objective of this panel is to discuss how undergraduate research can be successfully carried out in liberal arts schools in spite of all these constraints. We want to examine the mechanics of undergraduate research in such an environment -which practices work and which do not.The mechanics of undergraduate research include: How do we attract students to participate? How do we select students? What level of students? Do we train students in the research process? If so, what works? How do we select a topic appropriate for undergraduate research? Who does so? How do we circumscribe its scope? What kinds of activities are suitable as undergraduate research? When should students engage in research? Should we impose a research regimen on the students? If so, what works? How do we monitor the progress of research? What reporting mechanisms work? How do we keep students motivated? How do we encourage critical and independent thinking? How do we evaluate student research, and reward it? What are some avenues for disseminating student research? How do we train students to document their work, write up the results and present their research? What official mechanisms (courses/credits/internships, etc.) can we use to institutionalize research? How do we fund the research with or without external support? How can we ensure that the work done by the students can be reused and built upon once they leave? What can we do to build a research community among undergraduate students, complete with community spirit and structures for mutual support?The panelists represent a diversity of viewpoints: externally funded versus unfunded research, formal versus informal research regimen, independent versus course-based research, and modular versus ongoing research, to name a few. However, all the panelists are from small-to-medium-sized liberal arts colleges and have considerable experience supervising undergraduate research. In their statements, the panelists have listed six or more lessons that they have learned from experience. During the panel, the panelists will elaborate on these lessons and illustrate them with anecdotes. The panel will include time for the audience to share their experiences, and raise additional issues and concerns. We expect the panel to result in a well-rounded discussion of the mechanics of undergraduate research, and include multiple viewpoints and contrasting experience reports.

No abstract