CS 458/558 Project

Assigned:	Wednesday March 6
Proposal Due:	Monday April 8 (in canvas)
Deadline:	Thursday May 2, 11:59pm (last day of reading period for Yale College)

You must get your project approved by April 8th. Submit your proposal in Canvas. Submit your project on the zoo as hw 6.

As stated in the syllabus,

Each student will complete a final project comprising an automated decision system in a domain of interest. The assumption is that the student will use one of the techniques discussed in the course, and implement the system in Python. The instructor will provide a list of suggested topics, as well as entertain original proposals. Students may work in groups. The expectation is that a group project should have more substance than an individual project.

If you work in a group, your proposal and project needs to specify the division of labor - who is responsible for what. You should explain how the additional people in the group enable the project to be better - number of examples, size of domain, testing / quality assurance, user interface. Explain why more is better.

You may use late days when submitting the final project.

One of the grading criteria is your program's ability to explain its decision. A black box solution, such as a neural network, does NOT have this property. However, ways exist to provide explanations for them, e.g., local interpretability methods like mimic models, or counterfactual generation. (Sam can provide references if you are interested. There is a literature you can review and some interesting applications that one can come up with for them.)

Additional specifications:

Your project should run in the zoo. If you need additional libraries or software installed, you must clear it with the TF.
You need to include a write-up explaining your project, including how to run it, what it does, what is interesting or important about it, how it explains the decision. Screen shots might be helpful as well.
You can create a web site. In this case, it may not need to run in the zoo. However, you need to submit all the code so that it could theoretically be run in the zoo.
I prefer that you implement your project in a jupyter notebook in Python. However, you may use other languages and do not have to use jupyter.
Can you leverage a project for another course? It depends. You need to check with me first.
Can you change your project after your proposal? Well, yes. The real question is what happens if you do that? The answer is again, it depends. If your initial idea hits a wall and does not work, you probably should figure out another way to go.
During the last week of class, we will have in-class project presentations. They are mandatory for group projects and optional for individual projects. They are worth 5 bonus points.

CS 458/558 Project Suggestions

While you are free to come up with your own project, here are some suggestions.

VOTE: currently written in Common LISP. See /c/cs458/lisp
- Rewrite in Python
- Apply to other domain
- Develop taxonomy of decision strategies (and meta-strategies) for one or more domains.
- Add a new language. It already knows English and French.
Emotions applied to a decision system.
Recommendation system. Smart search, like a reference librarian who understands your goals.
Qualitative arithmetic (When is a number a "good number" or a "bad number"?)
Qualitative pre/post analysis for quantitative analysis. (Pre: Why perform test? Why use this data? / Post: what is the meaning of the results? what action should I take? should I perform additional tests?) Think of this as a statistician's assistant.
Option generation: if a or b, why not c? If you don't like either choice, come up with another option.
Create a Stakeholder taxonomy for specific decision domain. For example, you are designing a new electric vehicle. What do different stakeholders care about? Are there conflicts? Are there design strategies to mitigate those conflicts?
Build a rule base expert system. Or translate Norvig's mycin or emycin from Common Lisp to Python. See Paradigms of Artificial Intelligence Programming, Chapter 16.
Build a case based system. See Case-Based Reasoning: A Research Paradigm, Slade.
Finance application: equity analysis (expanded hw2), risk management, big data.
Example: Create a risk management system for equity portfolios. Given a portfolio of US or international stocks, across the capital spectrum (large, mid, small), create a system that will calculate a 5 day value at risk (VAR) at the 95% level. Also, perform stress tests and scenario analyses using both historical data and prospective scenarios. For international stocks, include foreign exchange (FX) risk analysis. The system should produce a summary heat map analysis (Green, Yellow, Red) indicating the risk level of the portfolio. Extra credit: if the portfolio is in the red range, the system can suggest trades that would mitigate the risk.
Monte Carlo simulation a la hw1
Machine Learning. Go for it, but add a layer of goals or relationships or explanations.

CS 558 requirement

If you are enrolled in the graduate version of the class, you need to include a review of the literature related to your project. We are talking annotated bibliography.