COMS 4995: Incentives in Computer Science


Course Assistants:

Time/location: 1:10-3:40 PM Wednesdays, in room 411 of the International Affairs building.

Piazza site: here

Prerequisites: Mathematical maturity at the level of a 4000-level course such as algorithms (CSOR 4231) or complexity (COMS 4236). Programming maturity at the level of COMS W3134.

Course Description: Many 21st-century computer science applications require the design of software or systems that interact with multiple self-interested participants. This course will provide students with the vocabulary and modeling tools to reason about such design problems. Emphasis will be on understanding basic economic and game theoretic concepts that are relevant across many application domains, and on case studies that demonstrate how to apply these concepts to real-world design problems. Topics include auctions (for spectrum and online advertising), equilibrium analysis, cryptocurrencies, network protocols, two-sided markets (online labor markets, dating markets, etc.), crowdsourcing, reputation systems, and social choice. Case studies include Bitcoin and Ethereum, eBay's reputation system, Facebook's advertising mechanism, Mechanical Turk, and Augur's prediction markets.

Textbook: There is no official textbook for the course. There will be lecture notes for most of the lectures. One general reference is Twenty Lectures on Algorithmic Game Theory, which is based on a more advanced graduate course. (The overlap with this course will be roughly 20-25%. Though if you enjoy this course, you're likely to also enjoy many of the topics in this book.)

Lecture notes


Tentative Syllabus (will likely change):

Detailed Lecture Schedule