CSC594 Topics in Artificial Intelligence:
Emotion and Personality Models for AI Agents
Professor Clark Elliott

Overview

Assignments:

• Follow the instructions at D2L to be added to the class Google Group.

• Follow the instructions in D2L | Content | Getting Started to create and share your Google Root Content Theory Document with the class Google Group.

• Post your short bio to the D2L Say Hello forum. Respond the three others. Make FIVE "expert" postings to the class forums relative to the content of the class. These can be new threads, or replies to existing threads. (The "Expert" designation just means they are relative to AI, Emotion Reasoning, Content Theories, or other content in the class.). Concatenate all your postings together and submit all nine to the D2L dropbox.

• Complete your Mini Personality Content Theory

• Demonstrate Socket Communication using strings. We want you to be able to communicate with your AI program via a minimal socket interface. This is proof-of-concept communication between two "agent" shells. I give you all the basic java code. You do not need to be more than an absolutely minimal Java programmer to complete this assignment. You just need to get the example running. If you prefer to work in another language, such as Python or C++, that is fine.

• Make regular progress on your Ideas File throughout the quarter, and complete it for review at the end of the quarter. Complete the Ideas File Admin quiz early and often.

• Select two research papers on AI models of emotion and personality either from the collection given in D2L or from another source. Read them. Write a summary paragraph. Write an opinion paragraph on why we would or would not want to read them. On your RootCT document: list, in order, the complete bibliographic reference for each paper, the summary and your opinion. Read the shared summary and opions of at least two others, and leave Google comments for them. Complete the Admin quiz for two papers early and often. Post a note on the forums that you are ready for comments, and include the direct URL to your share.

• For the bibliography and Structured Outlines assignments, complete:
  1. Part I Scraping Bibliographic Sources.
  2. Part II Citations and Bibliographies in Word using your personal, permanent bibliographic database.
  3. Part III Formal Structured Outline (with bibliography and citations added).

• Assignment: LISP. Run the sample LISP programs. Complete the following LISP AI programs. These will take some thought, but the actual programming is not hard, and I give you much of the code. You do not need a programming background to complete these assignments:

  • Plain-text editor recommendations: For casual programmers I recommend Atom / For AI scientists and advanced programmers I recommend spending the day to learn emacs. Note: You do not need to install and run the emacs editor to complete these small programming assignments. I give you alternative text editors you can use. However, emacs is an extremely powerful editor that runs everywhere, and it is the native environment for LISP. As part of the recorded tutorials, I make it available and give you a quick tour through its use.

  • Run the sample programs!

  • Complete the Towers of Hanoi assignment: Requires lots of thought, but only a couple lines of code to modify. For novice programmers 1-2 hours of thought to get it running the first time. Six minutes to get it running the second time.

  • Complete the animal learning program assignment, including the "advanced" adding of intensity. (Note: the thought is challenging, but the actual programming is not hard.) Complete the optional "very advanced" range of emotions if you are up to it.

  • Complete the LispPrograms Admin Quiz instead of the checklist. Zip together your Towers of Hanoi program, your Animal program, your Animal data file and your Animal console text file (copy and paste from the console showing your program running). Submit to D2L.

  • Extra credit (not required):

    • Complete the Bermudez Extra Credit quiz.

    • Complete the Neural Network Perceptron Logic Gates assignment. (Requires NN lecture by Elliott based on Bermudez)

    • Complete the Neural Network Perceptron Toy Face Recognition assignment. (Requires NN lecture by Elliott based on Bermudez)

    • Zip together your two programs (or don't zip if just one) and submit to D2L.

• The Affective Reasoner papers can be found here. Read Modeling the Emotions of Others and Emotion Intensity. Become familiar with the AR Emotions Table. Then read, or scan Affective Story-morphing of Frankenstein. By starting or responding to a D2L discussion thread, make at least three substantial posts on these topics. Complete the ThreeAR-Papers Admin quiz.

• Assignment: Simple Rules Programming assignment. Run the toy system. Write a few rules of your own to achieve some goal.

• Assignment: Story Morph Complete your own story morph and submit to D2L.

• Throughout the quarter, maintain your Google Content Theory Root Document so that you meet the progress requirements for the various checkpoints during the quarter.

Associated links:

• Some Affective Reasoning Project Threads
• Simple Logic
• Overview of topics, requirements, etc.
• Paper Critique Template
• Using Panopto to create a three-minute (or less) video of your running AI code.
• Structured Outlines description.

Extra Credit

Important general links:

• Submission File Formats
• Academic Integrity Sanctions. All grades are subject to Academic Integrity Rules. Make SURE you understand the rules for this class.
• Checklists for Programs.
• Visit DePaul D2L to see all documents, deadlines, and assignments.
• Submission & Late Assignments policies.
• Cheating is wrong.