Artificial Intelligence Methods for Social Good (Spring 2020)

Basic Information

Course Name: Artificial Intelligence Methods for Social Good
Meeting Days, Times, Location: Tue/Thu 10:30am-11:50am, GHC 4215
Semester: Spring, Year: 2020
Units: 9/12, Section(s): 17-737/17-537

 

Instructor Information

Name Dr. Fei Fang
Contact Info Email: feifang@cmu.edu
Office location Wean Hall 4126
Office hours Tue 3pm-4pm or by appointment

 

TA Information

Name Ryan Shi
Contact Info Email: ryanshi@cmu.edu
Office location Wean Hall 3130
Office hours Wed 4:30pm-5:30pm or by appointment

 

Course Description

The rapid advance in artificial intelligence (AI) has opened up new possibilities of using AI to tackle the most challenging societal problems today. This course brings together a set of advanced AI methods that allow us to address such challenges and promote social good. We will cover a wide range of topics in AI, including:

  • Search, Planning, and Optimization [OPT]: planning and scheduling, convex optimization, mathematical programming;
  • Multiagent Systems [MAS]: computational game theory, mechanism design, human behavior modeling;
  • Machine Learning [ML]: classification and regression, clustering, probabilistic graphical models, deep learning, reinforcement learning.

In addition to providing a deep understanding of these methods, the course will introduce which societal challenges they can tackle and how, in the areas of (i) healthcare, (ii) social welfare, (iii) security and privacy, (iv) environmental sustainability. The course will also cover special topics such as AI and Ethics.

The course content is designed to not have too much overlap with other AI courses offered at CMU. Although the course is listed within SCS, it should be of interest to students in several other departments, including ECE, EPP, and SDS. The students will work in groups on a project exploring the possibility of using AI to help address a societal problem, with a project report and poster presentation delivered at the end of the semester.

(9 Unit) The students in this 9-unit course are expected to have taken at least three mathematics courses covering linear algebra, calculus, and probability.

(12 Unit) This 12-unit course is only open to graduate students (master and Ph.D. students) with previous programming experience and background knowledge in artificial intelligence.

Please see the instructor if you are unsure whether your background is suitable for the course.

 

Learning Objectives

At the end of the course, the students should be able to

  • Identify societal challenges that can potentially be tackled by AI methods, and determine which AI methods can be applied
  • Describe the AI methods covered in the course, including the basic concepts, the key algorithms, and the commonly-used implementation of the methods
  • Model the societal challenges as mathematical problems that AI techniques can be applied and propose how to adjust and modify the AI techniques to fit the problems
  • Describe the evaluation criteria and methodologies of applying AI methods for social good
  • Deliver a report of course project and present the work through a poster presentation

 

Course Schedule (Subject to Change)

    Last update: 1/24/20

# Date Topic Slides and References
1 1/14 (Tue)   Introduction, Logistics, Course Project Slides: Available on Canvas

Artificial Intelligence for Social Good: A Survey

2 1/16 (Thu)   [OPT1]: Basics of Optimization

Cover: Convex optimization, Linear Programming (LP) and Mixed Integer Linear Programming (MILP)

Slides: Available on Canvas

Convex Optimization, Ch. 1-4

Applied Mathematical Programming, Ch. 2, 9

3 1/21 (Tue) [OPT2]: Conservation Planning

Cover: Wildlife corridor design

Slides: Available on Canvas

(PRA1) Solving Connected Subgraph Problems in Wildlife Conservation

Trade-offs and efficiencies in optimal budget-constrained multispecies corridor networks

Robust Network Design for Multispecies Conservation

4 1/23 (Thu) [ML1]: Basics of Regression and Classification

Cover: Linear and Logistic Regression, Kernel Regression, Decision Trees

Slides: Available on Canvas

Pattern Recognition and Machine Learning, Ch. 3, 4, 6

5 1/28 (Tue) [ML2]: Data-based Prediction

Cover: Food rescue, Detecting social bots on Twitter

(PRA2) Improving Efficiency of Volunteer-Based Food Rescue Operations

Botornot: A system to evaluate social bots

Scalable and generalizable social bot detection through data selection

6 1/30 (Thu) [MAS1]: Basics of Game Theory

Cover: Equilibrium concepts

Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations, Ch. 3-5
7 2/4 (Tue) [MAS2]: Security Games

Cover: Ferry protection, ranger patrol planning

 Deployed ARMOR Protection: The Application of a Game Theoretic Model for Security at the Los Angeles International Airport

Optimal Patrol Strategy for Protecting Moving Targets with Multiple Mobile Resources

Deploying PAWS: Field Optimization of the Protection Assistant for Wildlife Security

8 2/6 (Thu) [GL] Guest Lecture by Haifeng Xu

Cover: Bayesian Persuasion and Security Games

9 2/11 (Tue) [GL] Guest Lecture by Ding Zhao

Cover: Autonomous Driving

(PRA4) Accelerated Evaluation of Automated Vehicles Safety in Lane-Change Scenarios Based on Importance Sampling Techniques

Designing importance samplers to simulate machine learning predictors via optimization

10 2/13 (Thu) [ML3] Basics of Reinforcement Learning

Cover: Markov Decision Process (MDP), Q-Learning, Policy Gradient

Reinforcement Learning: An Introduction, Chapter 3, 6, 13

 

11 2/18 (Tue) [ML4] Reinforcement Learning for Bike Repositioning

Cover: Bike Sharing

Dynamic Bike Reposition: A Spatio-Temporal Reinforcement Learning Approach
12 2/20 (Thu) [GL] Guest Lecture by Hong Shen

Cover: AI and Ethics and Policy

13 2/25 (Tue) [ML5]: Basics of Deep Learning

Cover: Neural Networks, Object detection using Faster R-CNN

Deep Learning, Ch. 6, 9

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

SPOT Poachers in Action: Augmenting Conservation Drones with Automatic Detection in Near Real Time

14 2/27 (Thu) [ML6]: Learning from Remote Sensing Data

Cover: Poverty and crop yield prediction

(PRA) Combining satellite imagery and machine learning to predict poverty

Deep Gaussian Process for Crop Yield Prediction Based on Remote Sensing Data

15 3/3 (Tue) [OPT3] Basics of Influence Maximization

Cover: Influence propagation models, submodular function optimization

Maximizing the spread of influence through a social network

Submodular Functions: Extensions, Distributions, and Algorithms. A Survey

Information and Influence Propagation in Social Networks

16 3/5 (Thu) [OPT4] Dynamic Influence Maximization under Uncertainty

Cover: HIV prevention

(PRA) Using Social Networks to Aid Homeless Shelters: Dynamic Influence Maximization under Uncertainty

Influence Maximization in the Field: The Arduous Journey From Emerging to Deployed Application

Uncharted but not Uninfluenced: Influence Maximization with an Uncertain Network

17 3/17 (Tue) [MAS3]: Mechanism Design with Money
Cover: Auction, Truthfulness, Price-of-Anarchy
Algorithmic Game Theory, Ch. 9, 11
18 3/19 (Thu) [MAS4]: Building an Electronic Agricultural Market (PRA) Designing and Evolving an Electronic Agricultural Marketplace in Uganda
19 3/24 (Tue) [ML7]: Mixture Models and Probabilistic Graphical Models

Cover: Gaussian Mixture Models (GMMs), Dynamic Bayesian Networks (DBNs), Markov Random Fields (MRFs)

Pattern Recognition and Machine Learning, Ch. 8, 9
20 3/26 (Thu) [MAS5]: Human Behavior Modeling

Cover: Prospect theory, Quantal response

Improving Resource Allocation Strategy Against Human Adversaries in Security Games

Predicting human behavior in unrepeated, simultaneous-move games

21 3/31 (Tue) [MAS6]: Predict Human Activity in Adversarial Settings “A Game of Thrones”: When Human Behavior Models Compete in Repeated Stackelberg Security Games

Keeping pace with criminals: Designing patrol allocation against adaptive opportunistic criminals

Taking it for a Test Drive: A Hybrid Spatio-temporal Model for Wildlife Poaching Prediction Evaluated through a Controlled Field Test

Cloudy with a Chance of Poaching: Adversary Behavior Modeling and Forecasting with Real-World Poaching Data

CAPTURE: A New Predictive Anti-Poaching Tool for Wildlife Protection

22 4/2 (Thu) [OPT5]: Combinatorial Optimization and Robust Optimization

Cover: Duality, branch and price, maximin model, minimax regret, prevent illegal fishing

Combinatorial Optimization: Algorithms and Complexity, Ch. 3

Branch-and-price: Column generation for solving huge integer programs

Robust protection of fisheries with COmPASS

23 4/7 (Tue) [OPT6]: Optimizing Kidney Exchange

Cover: Kidney Exchange

FutureMatch: Combining Human Value Judgments and Machine Learning to Match in Dynamic Environments [Extended version]

Position-Indexed Formulations for Kidney Exchange [Extended version]

24 4/9 (Tue) [OPT6] MDP Planning

Cover: Partially Observable MDP, Monte-Carlo Planning

Planning and acting in partially observable stochastic domains

Monte-Carlo Planning in Large POMDPs

Bandit based Monte-Carlo Planning

25 4/14 (Thu) [GL]: Guest Lecture by Thomas Dietterich

Cover: Ecosystem Management, Invasive species management

(PRA) Simulator-Defined Markov Decision Processes: A Case Study in Managing Bio-invasions

PAC Optimal MDP Planning with Application to Invasive Species Management

PAC Optimal Planning for Invasive Species Management: Improved Exploration for Reinforcement Learning from Simulator-Defined MDPs

26 4/21 (Tue) [MAS6]: Spatio-Temporal Pricing in Ridesharing Platforms

Cover: scheduling and pricing, competitive equilibrium

Spatio-Temporal Pricing for Ridesharing Platforms

Hongyao Ma, Fei Fang, David C. Parkes

27 4/23 (Thu) Poster Presentation at CMU AI for Social Good Symposium

Location: GHC 6115

28 4/28 (Tue)
29 4/30 (Thu)
5/5 (Fri) Final Project Report due

 

Learning Resources

No formal textbook. References and additional resources will be provided in slides and on Canvas.

 

Assessments

The final course grade will be calculated using the following categories:

Assessment Percentage of Final Grade
Class participation 10 points
Paper Reading Assignment 20 points
Online Homework 20 points
Course Project 50 points

 

  • Class participation. The grading of the class participation will be mostly based on attendance, checked by in-class polls and asking and answering questions in class. Other factors include asking and answering questions on Canvas.
  • Paper reading assignment. The course will require all students to complete weekly paper reading assignments individually. In each assignment, the students are required to provide a summary of the paper/article,  questions, and discussion. The summary should cover “target problem”, “why is AI needed”, “intervention overview”, “data used”,  “resource needed”, “deployment status” (check Section 3.4.3 in the survey paper for reference). The discussion can be about potential improvement or future direction, or a brainstorming idea. The assignments will be submitted through Canvas and will be peer-reviewed, but the final score will be provided by the instructor and the TA.
  • Online homework. The course will require all students to complete biweekly online homework assignments individually. Each assignment will involve checking the understanding of basic concepts and working through the algorithms presented in class on example problems. Most questions are multiple-choice questions or numerical answer questions.
  • Course project. The students will work in small groups (2-3 students in each group) on a project exploring the possibility of using AI to help address a social good problem. The students are expected to focus on one or more societal challenges, propose models and AI-based solutions to tackle the challenges and evaluate the solutions. The students are required to submit a project report through Canvas and deliver a poster presentation. The progress of projects will be checked through the Project Proposal, Project Progress Report, Poster Presentation, and Final Project Report. The proposal and progress report will be peer-reviewed. The presentation and the final report will be evaluated by the instructor and TA directly.
    • For students in the 12-unit section, the final report will get a full score if it is at the same level as accepted papers at AAAI-20 AI for Social Impact track (see evaluation criteria here) or NeurIPS 2019 Joint Workshop on AI for Social Good (see accepted workshop papers for track 1 here).
    • For students in the 9-unit section, the final report will get a full score if it is comparable in quality to the following sample reports: [Sample report 1][Sample report 2]
    • For all the reports, students may choose a format with the guidance of their advisors or use the LNCS format or AAAI format.

Students will be assigned final letter grades according to the following table.

Grade Range of Points
A [90,100]
B [80,90)
C [70,80)
D [60,70)
R (F) [0,59)


Grading Policies

  • Late-work policy and Make-up work policy: No late days or make-up work is allowed. However, for both paper reading assignments and written answer assignments, the lowest scored assignment in each category will be dropped.
  • Re-grade policy: To request a re-grade, the student needs to write an email to the instructor titled “Re-grade request from [Student’s Full Name]” within one week of receiving the graded assignment.
  • Attendance and participation policy: Attendance and participation will be a graded component of the course. The grading of the class participation will be mostly based on attendance, checked by in-class quizzes and asking and answering questions in class. Other factors include asking and answering questions on Canvas.

 

Course Policies

  • Academic Integrity & Collaboration: For both paper reading assignments and written answer assignments, a student can discuss with other students, but he need to specify the names of the students he discussed with in the submission, and complete the calculations and writing of explanations, summary, and questions on his own. For the course project, the students can discuss and collaborate with others (including students, faculty members, and domain experts), but the students need to give proper credits to whoever involved, and report the contributions of each group member in the final report and presentations, which will be considered in the grading. For assignments and the course project, it is allowed to use publicly available code packages but the source of code package needs to be specified in the submission. Plagiarism is not allowed. The policy is motivated by CMU policy on academic integrity which can be found here.
  • Mobile Devices: Mobile devices are allowed in class. Cellphones should be in silent mode. Students who use tablet in upright position and laptops will be asked to sit in the back rows of the classroom.
  • Accommodations for students with disabilities: If you have a disability and require accommodations, please contact Catherine Getchell, Director of Disability Resources, 412-268-6121, getchell@cmu.edu. If you have an accommodations letter from the Disability Resources office, I encourage you to discuss your accommodations and needs with me as early in the semester as possible. I will work with you to ensure that accommodations are provided as appropriate.
  • Statement on student wellness: As a student, you may experience a range of challenges that can interfere with learning, such as strained relationships, increased anxiety, substance use, feeling down, difficulty concentrating and/or lack of motivation. These mental health concerns or stressful events may diminish your academic performance and/or reduce your ability to participate in daily activities. CMU services are available, and treatment does work. You can learn more about confidential mental health services available on campus here. Support is always available (24/7) from Counseling and Psychological Services: 412-268-2922.