LE/EECS 4401 3.0 Artificial Intelligence
GS/EECS 5326 3.0 Topics in Artificial Intelligence
Winter 2022

Course Project

This course requires doing a project in the area of artificial intelligence, on a topic related to the material covered in the course. The project may involve doing a literature survey, an evaluation of a an AI solver or tool, an implementation of an AI method, or even a research paper on an AI problem and a proposed solution. Some topics suggestions appear below, classified according to the type of project. Projects in EECS 4401 must be done in teams of three students; in EECS 5326, the project must be individual. For team projects, each student must be responsible for a distinct part of the project presentation and a distinct part of the final report, which will be graded separately (the introduction and conclusion of the report can be common).

Project Deliverables

On April 5 and 7 students will give a short (approx. 15 minutes) class presentation on their project and answer questions (you will be given a time slot in advance). The slides used in the presentation must be submitted in eClass by April 7 (in PDF or Powerpoint format). Every team member must submit the same presentation,, which should list all team members, and state which team member presented which part.

Your final report on the project is due by April 11 for EECS4401 and May 2 for EECS5326. For literature survey projects, include an overview of the topic area, a discussion of the papers read, and a complete list of references; critical analysis is expected. For system evaluation or implementation projects, state the project's motivation and objectives, provide a readable description the system that was implemented or evaluated, explain how the system is used, describe the testing that was performed, and discuss how usable the system is; include the documented code and/or sample tests in appendix as appropriate. Note that it is not sufficient to simply download a solver and run some of the examples included with it; you must either develop some new examples or evaluate the solver on a wide range of benchmarks. (Every team member must submit the same report, which should list all team members, and say which team member was responsible for which sections.)

Suggested Project Topics

• Experiment with and evaluate a solver/reasoner such as for example, the Z3 SMT solver, the Clingo ASP solver, etc.
• Investigate the description logic languages that are part of OWL and/or experiment with an ontology editor/reasoner platform such as Protege; examine how description logics can be used for ontology based data access.
• Investigate the PDDL planning domain specification language and/or experiment with planning systems (see the the Haslum et al. book as well as Muise's http://planning.domains for some domains, a classical planner, and a domain editor; Muise's site also provides other systems for FOND planning and multiagent epistemic planning).
• Experiment with an agent programming language such as Golog (see Reiter's book Ch 6, which also discusses a simple Prolog implementation with some examples; more complete implementations include Levesque's Ergo and Sardina and Vassos's IndiGolog) or Jason.
• Implement a reinforcement learning algorithm (e.g., Q-learning, SARSA) and experiment with it on some problems; there are libraries of problems such as OpenIA Gym; a recent development is Toro Icarte's reward machines (see Russell and Norvig, Ch. 22 and the references on the course web page).
• Survey work on large commonsense knowledge bases (or knowledge graphs) such as DBpedia, and WordNet, Cyc, etc.; some knowledge graphs are partly build using machine learning techniques; compare these approaches to those that use formal ontologies defined in OWL.
• Survey recent work on a topic such as AI safety, explainable AI, or deep reinforcement learning.