(Writing and Research in Liberal Arts Computer Science)
|Summary||The WARLACS AI assignments both give students practice applying a variety of AI ideas and scaffold the development of their writing and research skills, culminating in a final project that models the basic research process at a small scale.|
|Topics||The AI topics covered by the assignments are heuristic search, adversarial search, and reinforcement learning. Other related topics include communicating AI ideas to a variety of audiences, engaging with the AI literature, designing experiments, and the peer review process.|
|Audience||These assignments are intended for an undergraduate introduction to AI course.|
|Difficulty||These projects have open-ended problems requiring creativity and experimentation, which presents a challenge to most students at this level (some more than others). Each of these assignments is intended to take between 5 and 10 hours per week (some projects span multiple weeks). When possible, structures are in place to allow students to self-assess how well they have done so far, which may motivate them to ask for help or help them do an analysis of the costs and benefits of continuing to work on the project. The level of challenge can also be adjusted by providing more or less structure (e.g. in the form of hints).|
Software requirements: The projects use Python 3.x with Tkinter and are otherwise self-contained.
Knowledge requirements: To be successful, students should have general college-level writing instruction, comfort with making technical arguments (e.g. basic proof techniques), comfort with programming and computational problem solving (ideally using the Python language), and an introduction to each specific project topic (ideally soon before the project is assigned). To gain extra credit points, students should be familiar with intermediate-level algorithms content (e.g. minimum spanning tree algorithms) and/or comfort with discovering, learning about, and utilizing new algorithms/modules.
|Variants||There are many ways to adapt/re-use elements of these assignments:
In order to preserve the integrity of these assignments for other users, please do not distribute solutions or grading scripts for these projects.
These materials are available by request to instructors with verified professional emails. Please contact Erin J. Talvitie at Harvey Mudd College (firstname.lastname@example.org).
The WARLACS AI Assignments, particularly the first two projects, owe a great deal to inspiration from The Pac-Man Projects by John DeNero and Dan Klein, which were added to the Model Assignments archive in 2010 and remain popular in many introductory AI courses. These assignments have also been significantly shaped by constructive feedback from many undergraduate students at Franklin & Marshall College and Harvey Mudd College.
In addition to standard Python modules, these projects make use of the cImage.py module written by Bradley Miller. Project 1 contains a public domain image of an octopus from the Smithsonian Institute (source). Project 2 contains a public domain image of Connect Four (source).
This material is based upon work supported by the National Science Foundation under Grant No. IIS1939827. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.