10-709 Fall 2017: Fundamentals of Learning from the Crowd
Time: Tuesday and Thursday 1.30pm to 2.50pm
Location: GHC 4303
Units: 12
Instructor: Nihar Shah
Nihar's office hours: 3-4pm every Tuesday in GHC 8211
nihars at cs dot cmu dot edu
TA: Ritesh Noothigattu
Ritesh's office hours: 3-4pm every Thursday outside GHC 8013
riteshn at cmu dot edu
Description: Crowdsourcing is a burgeoning area that is popular in academic research, industrial applications, and also in societal causes. In this course, we will cover the foundational theoretical principles behind crowdsourcing and learning from the crowd. We will study this field via the lens of game theory (how to incentivize people to provide better data), learning theory (how to make sense of this data), and social choice theory (how to be fair). We will also touch upon literature in psychology and economics that studies the behavior of people. Along the way, we will discuss several fascinating paradoxes and conduct some live experiments in the class. Lectures will be taught on the board.
Evaluation: Homeworks, final project, class participation.
Prerequisites: Basic probability (e.g., the student should be comfortable with conditional expectations, the Gaussian distribution, union bound), basic linear algebra (e.g., singular value decomposition) and basic programming. All other required background in game thoery, learning theory, etc. will be taught in the lectures.
Class material: Will be posted on Autolab.
Tentative schedule (subject to change):
Sept 5 What is this course about? What is crowdsourcing?
Sept 7 How to win 40,000ドル?
Sept 12 How to properly make strict rules?
Sept 14 How do these rules help in machine learning?
Sept 19 How do casinos help in crowdsourcing?
Sept 21 Doubling down on casinos helping in crowdsourcing.
Sept 26 What was "A beautiful mind" all about?
Sept 28 When Thomas met John.
Oct 3 How to administer a virtual truth serum?
Oct 5 Some more serum for dessert please.
Oct 10 What are concentration inequalities? And no, its not related to Yoga.
Oct 12 What are good models for ranking?
Oct 17 How to rank in a simple, robust and optimal manner?
Oct 19 How to become more active?
Oct 24 How to grade your peers? (Guest lecture by Ritesh Noothigattu)
Oct 26 Who started the gossip?
Oct 31 I started the gossip. And you can't catch me. (Guest lecture by Prof. Giulia Fanti)
Nov 2 The labels are too crowded
Nov 7 Whats up with these parameters?
Nov 9 The world is a paradox.
Nov 14 We are not who we really think we are. Then who are we?
Nov 16 How can I be fair to everyone?
Nov 21 You cannot be fair to everyone.
Nov 28 Why won't you vote for your favorite candidate?
Nov 30 Finishing up. Do you have any questions?
Dec 5 Project presentations I
Dec 7 Project presentations II
Timeline for homeworks and project:
Project:
Sept 20, 1pm Group names
Oct 17, 1pm Proposal
Dec 5, 1pm Final report
Last two lectures Presentation
Major homeworks:
Sept 20, 1pm Major homework 1 release
Oct 8, 1pm Major homework 1 submission
Nov 3, 1pm Major homework 2 release
Nov 21, 1pm Major homework 2 submission
Final Projects: