Engineering Societal Systems / Spring 2025
Updates
Course Description
This seminar class will explore topics in the intersection of computer science, economics, and operations -- on the application of algorithms, data science/machine learning, and market design to the study of democracy, markets, and societal systems at large. This year, we will especially focus on the theme of computational systems in government, education, and related high-stakes settings, broken down into four kinds of systems: matching/recommendation, prediction, optimization, pricing. The goal is for the class to be methodologically appropriate for a diverse set of students, including in ORIE, CS, and IS. The assignments will have substantial flexibility for you to choose the methodological toolkit that is most appropriate for your research and goals. The paper discussions and in class lectures will also aim to be broadly accessible and interesting.
Important links
- Course website
- Syllabus
- Paper discussion signup
- Paper list
- Ed Discussion – Primary communication tool; automatically added from Canvas if registered. Let me know if not.
Course topics
- Introduction and overview
- What is market design? What are the tools used?
- Overview of classic papers/ideas
- How do we analyze existing systems? What is the role of an engineer/computing in understanding societal systems? How do we design new systems? Where has engineering of societal systems most succeeded? How do we choose an appropriate objective function? What are “wicked problems” and failure modes?
- Matching and recommendation
- School choice (matching + recommendations)
- Refugee matching
- Prediction for decision-making
- Prediction in admissions and student performance tracking
- Criminal risk assessments, crime prediction, prediction in child welfare
- Limits of prediction and “against predictive optimization”
- Optimization and mechanisms
- Gerrymandering, sortition
- Social service optimization (homeless services, medical resources, etc)
- Public transportation: school bus routing, stop placement, shared vehicles, school zone design
- Pricing
- Congestion pricing, wireless spectrum auctions, tax auditing policies
- Algorithmic pricing collusion
- Miscellaneous
- Algorithms in NYC government as an extended case study
- Experimentation in government
- Human-AI collaboration
- Limits of technical approaches
- Voting and social choice
- Other applications, such as online marketplaces
