Introduction to Algorithmic Economics

Introduction to Algorithmic Economics
SS/CS 149
Fall 2013

Academic Integrity

This course will require you to present and build on many existing sources. Proper attribution of ideas, text, and paraphrased text is required, in both oral and written presentation.

Homeworks

Homework assignments are designed to help you check that you are understanding the main concepts covered in class. You are strongly encouraged to collaborate with your classmates on homework problems, but each person must write up the final solutions individually. You should note on your homework specifically which problems were a collaborative effort and with whom. You may not search online for solutions, but if you do use research papers or other sources in your solutions, you must cite them.

Late policy: I will allocate each student 2 tokens at the beginning of the term. Each of these tokens can be used to buy a 24-hour extension on either homework during the term (you may spend them both on the same homework assignment). You should write clearly on your homework that you are using a token, and how many you are using. You cannot get extra tokens and zero credit will be given to late assignments.

Checkpoints

Many classes will include a 5-minute "checkpoint" exercise in class, to evaluate understanding. These are graded on either a 0/1 or 0/1/2 scale, and there is no makeup available.

Teaching

The instructor will offer a set of core course topics that will be taught by students in the class over the course of the semester. Each student must teach one session during the semester. When it is your turn to teach, your responsibilities are:

Five days before the class meeting you are scheduled to teach, email the instructor an outline of your session.
Design and conduct a 5-minute pass/fail self-assessment for your classmates to complete at the start of the next class meeting. This should be designed help your classmates determine whether they retained the most important points of your presentation.
By one week after you teach, email the instructor a brief (one- to two-paragraph) self-assessment of your teaching session. This should be based in part on the results of the assessment you designed.

Reaction Papers

The reaction paper is an opportunity to read two (or more) related papers in depth, reflect on them, and explore opportunities to build on them. You may work in a group of up to three people (larger groups should plan to cover more material in more depth). Together, you should choose and carefully read at least two papers not directly covered by the course. You should then (either together with your group members or individually) prepare a paper of 4-6 pages. Your paper should address the following:

Briefly summarize the contributions of the papers, taking care to relate the papers to each other. Why are these contributions interesting or important?
What are the weaknesses of the papers? How might they be addressed?
What are the most promising directions for extensions and future work in the area studied by the papers? How have (or haven't) they been explored in subsequent work?
Do some initial exploration of at least one of the possible directions for future work, and report on your successes and failures.

During the final week of class, each group will give a brief presentation on their reaction paper topic.

The reaction paper proposal should be approximately one page. It should describe the topic in a few sentences, suggest (a superset of) the papers to be read, list the group members, provide a one-paragraph summary of context for the topic (why is it important and interesting?), and provide a work plan (e.g., "In the first week, Alice will read paper 1 and Bob will read paper 2. We'll have a two-hour meeting to discuss. In the second week, ..."). Proposals should be submitted by email directly to the professor.

Resources

Seminars and workshops: email jenny AT hss DOT caltech DOT edu to be added to the Social and Information Science Lab (SISL) mailing list, to get announcements for weekly lunch talks (noon on Fridays).

The sigecom exchanges archives are a great place to start looking for ideas for your reaction papers. These articles provide a short, accessible introduction to a topic (they should not be one of the main papers you base your reaction paper on, but they'll give good references to other papers).

Conferences with mostly AGT papers:

ACM Conference on Electronic Commerce (EC): 2013, 2012, 2011, 2010, 2009, 2008, etc.
Workshop on Internet and Network Economics (WINE): 2013, 2012, 2011, 2010, 2009, etc.

Good CS Theory Conferences with at least a few AGT papers per year:

IEEE Symposium on Foundations of Computer Science (FOCS)
ACM Symposium on Theory of Computing (STOC)
ACM-SIAM Symposium on Discrete Algorithms (SODA)

Journal most likely to publish AGT work: