Computational Economics(Algorithmic Game Theory)

计算经济学

Xiang-Yang Li， Lan ZhangSchool of Computer Science and Technology, USTC

Professor, CS Department2021

» Course Logistics

» Instructors

» Course Basic Content

Today’s Class

2

3

Course Logistics

» Course Name: Computational Economics

» Instructor:• Xiang-Yang Li 李向阳

• Lan Zhang 张兰

» Class-Place:• 线下：中国科大高新校区智能楼 A202 :3(11,12,13) -- 19：

30～20：15; 20：20～21：05; 21：10～21：55

• 线上：

» Class-time：Wednesday, Weeks of 1 -16

» Course WeChat Group for discussion

Course Information

4

» Seminar-Type Course• Students will read research papers and survey

them

• Defense and Offense in discussions

» Hands-on Course• Two written homework sets

• Coding project covering the entire semester

• Survey of special topics

Requirements

5

» Class attendance (10%)

attendance 5%

active and asking questions 5%

» Two small homework sets (20%) (one homework set 10%)

» Each team of at most 3 students do the following projects

» One term survey paper (30%) (Each team expected to read papers from a chosen topic and write a survey/summary on this topic.)

» One term programming project (40%): This project is about programming and implementation related to computational economics:

1) project preparation and presentation (10%) (10 mins per group)

2) final project report (15%) (the project report need to report your methods, novelty, and results. Report need to follow IEEE/ACM conference format, and has at least 10 pages.)

3) project code and demo (15%) (at end of semester, demo to TA)

Grading System

6

One term survey paper (30%) (Each team is expected to read SEVERAL papers from a chosen topic and be able to write a survey about the topic)

» The selection of the topic/paper from the list is first-come-first-service (or you propose topics). No TWO students are allowed to select the SAME topic. You cannot COPY any material from any segment of results written by others (online material or published books, papers, reports), unless you need to cite some results or statements and clearly indicate in your report.

» The survey paper should be due at the end of the semester. The main content of paper (excluding the references) should be at least of 10 pages and in IEEE conference format, double column, font size (at most) 10,

» You can write it in Chinese or English. Need to have enough figures, formulas, citations (do not count as page limits)

1: Term Survey Paper

7

» Games for cooperative multi-agents

» Games for non-cooperative multi-agents

» Mechanisms for data collection crowdsourcing

» Incentive mechanisms for federated learning

» Mechanisms for combinatorial auctions

» Privacy-preserving mechanism design

» Deep multi-agent reinforcement learning

» Games for resource usage in cloud computing

» Distributed mechanism design

» Data quality evaluation with mutli-agents

» Online advertising

» Data trading and sharing

Some possible topics for survey

8

» Incentives for Mobile Crowd Sensing: A Survey

» Survey on Real Time Bidding Advertising

» A Survey of Incentive Mechanism Design for Federated Learning

Survey Papers Example

9

One term project, presentation, and report ( 40%)

» We provide a list of possible projects. You will select one project. AT most two groups can select the same project. You can also propose aproject and ask for our approval.

» You have to really implement the project and show that it works.

» Each group (of at most 3 students) needs to discuss the term project with the instructor or TA within 3 weeks of the first lecture. Each group needs to submit a 2 page project proposal by the end of the 3rd week. Submit your e-copy to TA, and hard-copy also to TA.

» Each student in the group will be graded equally unless it was reported to me and confirmed that some student did not do sufficient work for the project.

» Each group needs to do one 10 min presentations, at the end of the semester to demo the final results of your project.

2: Term programming Project

10

Although project areas must be approved, the choice is left largely up to the student. Projects may be theoretical or experimental.» The secretary recommendation problem

» matching algorithms for blood donation

» Pricing policies for data monetization

» Dynamic pricing with limited supply

» Strategies for dynamic data-driven decisions

» Task assignment in Crowdsourcing system

» Truth inference in Crowdsourcing system

» ……

Some examples of projects

11

» Pick a problem that is intellectually interesting, important, and improves the practice.

» 更加重要的是：你们能够完成，别人感兴趣

Choose Projects

12

» Question old school assumptions

» 挑战固有的假设

» Open your heart and mind to people who question assumptions

Look for blind spots

13

Components of your proposal

» Your project title

» Team members ( of at most 3 students)

» Challenges in your project

» Relation to the topic: computational economics issues

» Current literature on this topic

» What are the data? How to get the data?

» Your evaluation plan and metrics (how to evaluate the success of your project)

» Management aspects such as your project plan, critical paths, means of team communication (e-mail, chat room, meetings, version control system).

2.1: Project Proposal

14

Covers the following material:

» Explain your design.

» Discuss design alternatives,

» Summarize privacy or economics aspects of your project,

» Your system architecture.

» Your method or algorithms,

» Data to show the performance of your systems,

» The challenges faced by your group in implementing the project and how you address these challenges;

» Lessons learned from the project, and

» future plan for the project.

2.2: Project Presentation

15

Presentation

16

Presentation

17

Project report: Covers the following material: (at least 8 pages, IEEE Format， font size at most 10, two columns)

» Abstract, Introduction, ….

» Your system architecture: Explain your design, and compare with the literature; Discuss design alternatives,

» Your detailed method or algorithms and the technical challenges and how you address these challenges;

» Performance Data to show the performance of your systems,

» Discussion and Conclusion,

» References

2.3: Project Report

18

Project:

» Project Proposal: in three weeks (3 students per group)

» Project Presentation: three weeks before the end of semester (10 mins per group)

» Project Code and Demo: end of semester (to TA)

» Project Report: end of semester (last week, to TA)

» Survey Paper: end of semester (To TA, need e-copy)

Important Dates to submit coursework

19

» Each of you is expected to contribute to each class session by arriving on time, being attentive, participating in the class discussion if needed, and being respectful to your instructor and fellow students.

• Disruptive conversations, eating, sleeping and putting your feet on the furniture are not acceptable behavior in the class environment.

» In addition to arriving on time, students are expected to stay the whole class period.

• Please avoid disrupting fellow students and the instructor by arriving late or leaving early. If a situation arises that consistently causes you to be late or absent, please contact me.

» Every electronic device (anything with an on/off button) should be off during the class (exception: disability-helping devices).

Classroom Policy

20

21

Instructors Introduction

Who I am and What I do

22

Prof. XiangYang Li李向阳

http://staff.ustc.edu.cn/~xiangyangli

xiangyangli@ustc.edu.cnxiangyang.li@gmail.com

ACM Fellow, IEEE Fellow, ACM Distinguished Scientist

School of Computer Science and TechnologyUSTC

» Info

• Professor, School of Computer Science and Technology, USTC

• Professor, CS Department, IIT

• IEEE Fellow, ACM Distinguished Scientist 2015

• PhD/MS from UIUC 1997-2000

• BS, BE Tsinghua University 1990-1995

» Research Interest

• Wireless networks, mobile computing

• Big data security and privacy

• Internet of Things, Cyber Physical Systems

• Algorithm design and analysis, Game theory

» Supported by MIST, NSF, NSF China, RGC HongKong

» Contact Information

• Email: xiangyang.li@gmail.com

• 计算机学院，电三楼6楼 627，中国科学技术⼤学

Instructor

23

» ACM Fellow (2020), IEEE Fellow (2015)

» ACM Distinguished Scientist (2015)

» Best Paper/Demo Awards (10) or Candidates (2)• IEEE GlobeCom 2015 Best Paper

• ACM MOBICOM 2014 Best Paper

• ACM MOBICOM 2012 Best Demo

• ACM MOBICOM 2008 Best Paper Candidate (1 of 3)

• ACM MOBICOM 2005 Best Student Paper Candidate (1 out of 3)

• IEEE IPCCC 2014 Best Paper

• COCOON 2001 Best Paper

• IEEE HICSS 2002 Best Paper

» 2011 IIT, Research Excellence Award, School of Science

Awards and Honors

24

Who I am and What I do

25

Lan Zhang张兰

zhanglan@ustc.edu.cn

School of Computer Science and TechnologyUSTC

» Info

• Researcher, School of Computer Science and Technology, USTC

• Poster Doctor, Tsinghua University 2014-2016

• PhD, Tsinghua University 2007-2014

• BS, Tsinghua University 2003-2007

» Research Interest

• Big data understanding, protection and trading

• Mobile computing

» Award

• 2015 ACM China Doctoral Dissertation Award (1/2 nationally)

• CCF Outstanding Doctoral Dissertation Award (1/10 nationally)

» Contact Information

• Email: zhanglan@ustc.edu.cn

• 科技实验楼123，中国科学技术⼤学

Instructor

26

27

Course Basic Content

» What is computational economics?• Using computational thinking on economics

• Using economics thinking on computational problems

» Game theory• Basic concepts and computational issues• Nash Equilibria (computation and complexity)• Learning and Regret Minimization• Graphical Games, Cryptography and Game Theory

» Mechanism Design• Introduction, Mechanism with/without money• Computationally efficient approximation mechanisms• Profit/ social efficiency maximization, profit sharing• Combinatorial/ online/ distributed mechanism design• Preference expression

» Application in Computational Economics

What you will learn

28

Learning Objectives

29

» Learn the classic and state-of-the-art computational economics

» Study various applications where computational economics is needed and can be applied

» Challenge existing solutions and identify new problems in computational economics

Some expectations

30

» Participate in class, think critically, ask questions

» Read and write reviews critically

» Start on assignments and projects early

» Enjoy the class!

31

Some useful references

» Algorithmic Game Theory, Edited by Noam Nisan,TimRoughgarden, Eva Tardos, Vijay V. Vazirani,

• https://www.cs.cmu.edu/~sandholm/cs15-892F13/algorithmic-game-theory.pdf

» Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations

• http://www.masfoundations.org/mas.pdf

» Networks, Crowds, and Markets: Reasoning about a Highly Connected World by David Easley, Jon Kleinberg

• https://www.academia.edu/download/34741809/networks_crowds_and_markets.pdf

» ACM EC conferences

Some useful references

32

» http://www.cs.jhu.edu/~mdinitz/classes/AGT/Spring2020/

» http://cs.brown.edu/courses/csci1440/labs/

» http://jasonhartline.com/MDnA/ Mechanism Design and Approximation By Jason D. Hartline

» https://ml2.inf.ethz.ch/courses/agt/

» https://sites.northwestern.edu/hartline/tutorial-foundations-of-non-truthful-mechanism-design/ EC 2020 Tutorial: Foundations of Non-truthful Mechanism Design

» http://www.cs.cmu.edu/~arielpro/15896/schedule.html

» https://www.cis.upenn.edu/~aaroth/courses/agtS20.html

» http://www.eecs.harvard.edu/cs286r/courses/fall08/schedule.html

Some related courses

33

Any questions?

You can find me at:

» xiangyangli@ustc.edu.cn

» zhanglan@ustc.edu.cn

THANKS!