People make dozens of decisions every day in their personal and professional lives. What would it mean for you to trust a computer to make those decisions for you? It is likely that many of those decisions are already informed, mediated, or even made by computer systems. Explicit examples include dating sites like match.com or recommendation systems such as Amazon or Netflix. Most Internet ads on sites like Google or Facebook are run by real time bidding (RTB) systems that conduct split second auctions in the hopes of getting your attention. Driverless cars offer the promise of safer highways. Corporations and other enterprises invest in decision support systems to improve the quality of their products and services.
The spectrum of domains is diverse and impressive. As these systems become more accomplished, they also become more dissimilar. Like medicine, where most diagnostic tools and treatments rely on specialized knowledge, decision systems exhibit similar Balkanization. It is said that an expert is someone who knows more and more about less and less until they ultimately know everything about nothing. There is no grand, unified theory of disease. (Actually, there are researchers at Yale who propose inflammatory processes as such a theory.)
This course takes a meta approach. We will explore a cognitive process model of decisions that can be applied to almost any domain. We will simulate human decision making, based on goals, relationships, and emotions. We will apply this model to tasks such as preference, choice, explanation, planning, advice, and persuasion. Domains will include politics, ethics, finance, and technology.
We will have a number of guest speakers who will talk about real world systems and applications. Past speakers have come from the world of finance, as well as Google, Facebook, and Palantir. They are generally interested in recruiting as well. We often are able to arrange for a handful of students to have dinner with the speakers after class.
This work presents a goal-based model of decision making in which the relative priorities of goals drive the decision process -- a psychological alternative to traditional decision analysis. Building on the work of Schank and Abelson, the author uses goals as the basis for a model of interpersonal relations which permits decisions to incorporate personal and adopted goals in a uniform manner. The theory is modelled on the VOTE computer program which simulates Congressional roll-call voting decisions.
If you need help writing programs in Python 3, or want to update older Python 2 code, this book is just the ticket. Packed with practical recipes written and tested with Python 3.3, this unique cookbook is for experienced Python programmers who want to focus on modern tools and idioms
It's easy to start writing code with Python: that's why the language is so immensely popular. However, Python has unique strengths, charms, and expressivity that can be hard to grasp at first -- as well as hidden pitfalls that can easily trip you up if you aren't aware of them. Effective Python will help you harness the full power of Python to write exceptionally robust, efficient, maintainable, and well-performing code. Utilizing the concise, scenario-driven style pioneered in Scott Meyers's best-selling Effective C++, Brett Slatkin brings together 59 Python best practices, tips, shortcuts, and realistic code examples from expert programmers.
The course requirements consist of class attendance (required for guest speakers), several programming assignments in Python or Jupyter Notebooks, and occasional written homework and canvas quizzes, a paper and a final project. The programming assignments are an integral part of the course. For the paper, you will be asked to use Chat-GPT for brainstorming.
The lectures will be recorded. Attendance is required for guest speakers. On other days, there may be in-class quizzes to reward attendance.
Complete the online student information sheet.
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.
Late work without a Dean's excuse will be assessed a penalty of 5 points per day, based on the day and time recorded by the Zoo electronic submit program. At the end of term, up to 50 points will be deducted from the total lateness penalties your homework has accrued. However, according to Yale College regulations, *no* homework can be accepted after the end of Reading Week without a Temporary Incomplete (TI) authorized by your dean.
If you have a Dean's excuse or a TI, making up missed work may involve alternative assignments, at the discretion of the instructor; please check with the instructor in this case.
Unless otherwise specified, the homework assignments are your individual responsibility. Plagiarism is a violation of University rules and will not be tolerated. You must neither copy work from others (at Yale or elsewhere) nor allow your own work to be copied. You are definitely on the wrong side of the boundary if you give or receive a printed or electronic copy of your or anyone else's work for the course from this term.
You are encouraged to ask others for help with the computers and Unix, with questions about Python, general questions about the concepts and material of the course, but if you need more extensive help with a program or other assignment, please ask a TA or the instructor for assistance. Working in groups to solve homework problems is not permitted in this course, except for the final project. Please talk to the instructor if you have any questions about this policy.
During my years at Yale, I have had students who were either blind or deaf. They were generally among the best in the class. I want to make you succeed.
| Week | Date | Topic | Speaker | |
|---|---|---|---|---|
| 1 | Jan 17, 19 | Introduction and Overview of Decision Making | ||
| 2 | Jan 22, 24 | Economic Decision Theory. Behavioral Economics |
||
| 3 | Jan 29, 31 | AI, Cognitive Science, and Consciousness | ||
| 4 | Feb 5, 7 |
Feb 5: Guest speaker: John Niccolai, Citadel.
Goals Capital Budgeting / Net Present Value |
. | |
| 5 | Feb 12, 14 | Feb 12: Guest speakers: Joanne Lipman and Rebecca Distler. AI and the media. | . | |
| 6 | Feb 19, 21 | Feb 19: Guest speaker: Luciano Floridi, Director Yale Digital Ethics Center. Topic: AI and Ethics.
Feb 21: Guest speaker: Duke Dukelis, Google. Topic: Internet Ad Technology. |
||
| 7 | Feb 26, 28 |
Relationships, Explanations, Emotions Case based reasoning.
Quantitative Finance Feb 28: Guest speaker: Professor William Goetzmann, Yale School of Management. Finance. |
||
| 8 | Mar 4, 6 |
Mar 4: Guest speaker: Eren Orbey, Microsoft. Expert Systems |
||
| Mar 9-24 | Spring Break | |||
| 9 | Mar 25, 27 | Qualitative Arithmetic | . | |
| 10 | Apr 1, 3 |
April 1: Guest Speaker: Richard Apostolik, Global Association of Risk Professionals (GARP)< Risk Management as Exception Handling |
||
| 11 | Apr 8, 10 | Teleology of Technology (DWIM) | . | |
| 12 | Apr 15, 17 | Projects
April 17: Guest Speaker: Alborz Geramifard, Meta. Remote lecture over zoom. |
||
| 13 | Apr 22, 24 | Projects | ||