Artificial Intelligence meets Economics

News

• I have set up the ILIAS system for the seminar. Please take a look at it. There's also a forum and a Wiki. 

The main goal of the seminar is to provide the students with an advanced overview of the state-of-the-art principles and methods of artificial intelligence as they pertain to selected research topics in the fields of economics. Contrary to popular belief, AI and Economics have various problems in common  and, therefore, methodologies developed in AI have the potential to gain new insights when applied within economics and vice versa. Central to both fields is rational decision making, that is, given a  performance measure (a measure of  utility in the agent setting) how to derive the best course of action in an uncertain environment.

Specific topics that will be covered are:

1. Reasoning and decision making under uncertainty
2. Game theory
3. Statistical Relation Learning
4. Knowledge Management and Economics

During the first couple of meetings, the instructor will give a broad picture of the two fields and attempt to answer the following questions:

1. What are the problems and methods in both fields?
2. Where do they intersect?
3. What can one field learn from the other?

In particular, we will look at approaches to simple and complex decision making. Students will be assigned a seminar topic that fits their interest and existing knowledge. Both computer science and economics students are welcome.

Students are expected to write a seminar paper about a selected topic and to present their work to the other participants of the seminar. The seminar will meet bi-weeky.

Selected literature:

• Multiagent systems: Algorithmic, Game-Theoretic, and Logical Foundations. Yoav Shoham, Kevin Leyton-Brown, 2009.  (available as e-book)
• Stuart Russel and Peter Norvig: Artificial Intelligence - A Modern Approach, Prentice Hall 2003; http://aima.cs.berkeley.edu/ (2nd or 3rd edition)

Project examples:

• Installing a prediction market and writing  about the experience and results
• Setting up a simple auction system and writing up the experiences and results
• Writing about different methods to set-up a (keyword) auction market and contrasting their (dis-)advantages
• Using the Twitter API to extract mentions of a company and relating it to change in stock price
• Developing an agent for a game (poker, game theoretic game etc)
• Collecting literature about a particular game and the AI agents developed
• Your idea here...

Links and References:

• Iowa electronic markets (prediction market run by the University of Iowa)
• Zocalo is a toolkit for building prediction markets, markets in securities that pay out depending on outcomes of future events. They provide estimates of the likelihood of specific outcomes that are more reliable than other sources of predictions
• How to run a silent auction using Drupal
• Computer Poker Research Group at the University of Alberta
• Any suggestions? Post them in ILIAS

Location: B6, 26 Room A.206

Time:  13:45-15:15

• First meeting, Introduction: 16.2. (slides)
• Second meeting, Basics:  2.3.
• Third Meeting

Contact Dr. Mathias Niepert for details