CS3710 Advanced Topics in AI: AI for Social Good (Spring 2024)

Class time: Tuesday & Thursday 3:00pm – 4:15pm
Class location: Sennott Square 6516
Instructor: Ryan Shi
Email: ryanshi@pitt.edu
Office location: Sennott Square 5415
Office hours: By appointment on Calendly. If the times don't work for you, please send me an email with [CS3710] in the title.

Course Description

The rapid advance in 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 societal challenges, such as:

Optimization: linear and integer programming, influence maximization
Game Theory and Mechanism Design: security games, human behavior modeling, auction and market equilibrium
Sequential Decision Making: Markov decision processes, online planning, reinforcement learning
Natural Language Processing: large language models
Causal inference and experiment design
Algorithmic fairness

Armed with these tools, the course will introduce which societal challenges they can tackle and how, in areas such as healthcare, social welfare, security and privacy, and environmental sustainability. The course will critically examine the success stories in AI for social good, and explore how to enact real social change with AI.
Prerequisites: Working knowledge of probability and linear algebra, and a basic knowledge in AI. Although the course is listed within CS, it should be of interest to students in several other departments. Please contact the instructor if you are unsure whether your background is suitable for the course.

Textbooks

The course will not use any particular textbook. All reference materials will be provided on the course schedule page. We recommend the following books and articles for students who wish to explore AI for Social Good at a deeper level.

Geek Heresy. Kentaro Toyama, 2015.
Trustworthy Machine Learning. Kush R. Varshney, 2022.
The lone fan who became club president – and led his team to silverware. Jack Lang, 2023.

Case Study Presentations

We will have 13 case study sessions. Each case study is focued on an example AI4SG project. The instructor will lead the first session. For each of the remaining 12 sessions, we will have three designated student presenters in each of the three roles: author1, author2, and freestyler.

Author1 [20 minutes] You will be giving a presentation of Paper 1 as if you were the author. Treat this more as a tutorial presentation than a conference talk. You will want to go into the technical details and make sure your classmates understand the in's and out's of the paper.

Author2 [20 minutes]You will be giving a presentation of Paper 2 as if you were the author. Requirements for the talk are the same as the role of Author 1.

Evaluator [20 minutes] Choose one of the two papers, and do the following:

[10 minutes] Suppose this paper hasn't been published yet and you are the reviewer of this paper at a top conference. Complete and present a technical review of this paper. Then, take a holistic view of this project beyond the technical level, discuss the virtue and shortcomings of this project in the real world.

[10 minutes] For the second part of your presentation, you may pick one of the following options:

Situate this paper in the literature. Find and report on at least one older paper that substantially influenced the current paper and at least one newer paper that is influenced by the current paper.

Propose two follow-up projects to this paper: one in the same application domain but advances on technical side, the other using similar technical tools to a different application domain. Make sure to justify your proposal.

[Choose this only if you actually hope to involve a community organization in similar application domain in your course project] Explore the space of potential partners, reach out to them about potential AI use in their organization, and report on your findings.

You are required to present twice throughout the semester. The assignment is first come, first serve. Put down your names early on this spreadsheet! If no one signed up for the presentations 1 week before the class, I will randomly choose presenters from the 3 least active people (who signed up the least number of presentations on the sheet).

If you are not presenting, you are required to submit a case study writing assignment. This is typically a project summary of the papers studied in class. A template will be provided. Alternatively, you may choose to scribe class discussions during and after the presentations and submit the class minutes instead of a paper reading assignment. At most 3 students may choose to scribe for each class. You may scribe up to 3 times throughout the semester. Please also sign up for the scribe role on the spreadsheet ahead of time.

You are allowed to skip up to 2 case study writing assignments.

Case study writing assignments are due on Canvas one week after the class. Both paper summaries and discussion minutes will be shared with the class afterwards.

Course project

You will work on a research project exploring the possibility of using AI to help address a social good problem. You are expected to focus on one or more societal challenges, propose models and AI-based solutions to tackle the challenges and evaluate the solutions. You are encouraged to choose a topic that aligns with your own research. The project is meant to lead to a conference publication later. The project final report should be at a level acceptable to a workshop at a top-tier conference (see example papers here)

You may work individually or as a team of two students.

The progress of projects will be checked through the Project Proposal, Project Progress Report, Oral Presentation, and Final Project Report. The proposal and progress report will be peer-reviewed and then reviewed by the instructor. The presentation and the final report will be evaluated by the instructor directly.

Proposal due: February 2nd
Progress report due: March 8th
Oral presentations: April 18th
Final report due: April 26th

Course Schedule (Subject to Change)

#	Date	Topic	References
1	1/9	Introduction, Logistics Overview of AI for Social Good	Artificial Intelligence for Social Good: A Survey
2	1/11	Technical topic 1: (OPT) mathematical programming	Applied Mathematical Programming, Ch. 2 & 9
3	1/16	Case study 1: AI for food rescue	[Paper 1] Improving Efficiency of Volunteer-Based Food Rescue Operations [Paper 2] A Recommender System for Crowdsourcing Food Rescue Platforms
4	1/18	Meta topic 1: scoping an AI4SG project	Data Science Project Scoping Guide
5	1/23	Technical topic 2: (MAS) game theory	Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations, Ch. 3-5
6	1/25	Case study 2: security game for public safety	[Paper 1] Deployed ARMOR Protection: The Application of a Game Theoretic Model for Security at the Los Angeles International Airport [Paper 2] Optimal Patrol Strategy for Protecting Moving Targets with Multiple Mobile Resources
7	1/30	Case study 3: security game for wildlife conservation	[Paper 1] When Security Games Go Green: Designing Defender Strategies to Prevent Poaching and Illegal Fishing [Paper 2] Deploying PAWS: Field Optimization of the Protection Assistant for Wildlife Security
8	2/1	Technical topic 3: (OPT) Influence maximization	Information and Influence Propagation in Social Networks Maximizing the spread of influence through a social network Submodular Functions: Extensions, Distributions, and Algorithms. A Survey
9	2/6	Case study 4: AI for HIV prevention	[Paper 1] Using Social Networks to Aid Homeless Shelters: Dynamic Influence Maximization under Uncertainty [Paper 2] Clinical Trial of an AI-Augmented Intervention for HIV Prevention in Youth Experiencing Homelessness [Optional Reading] Influence Maximization in the Field: The Arduous Journey From Emerging to Deployed Application [Optional Reading] Uncharted but not Uninfluenced: Influence Maximization with an Uncertain Network
10	2/8	Technical topic 4: (RL) sequential decision making	Introduction to Multi-Armed Bandits Bandit Data-Driven Optimization for Crowdsourcing Food Rescue Platforms A contextual-bandit approach to personalized news article recommendation Reinforcement Learning: An Introduction, Ch. 3, 6, 13
11	2/13	Case study 5: Restless multi-armed bandit for public health	[Paper 1] Field Study in Deploying Restless Multi-Armed Bandits: Assisting Non-Profits in Improving Maternal & Child Health [Paper 2] Scalable Decision-Focused Learning in Restless Multi-Armed Bandits with Application to Maternal and Child Health
12	2/15	Guest lecture - Rex Chen
13	2/20	Case study 6: AI for bike repositioning	[Paper 1] Dynamic Bike Reposition: A Spatio-Temporal Reinforcement Learning Approach [Paper 2] Multi-Agent Reinforcement Learning for Order-dispatching via Order-Vehicle Distribution Matching
14	2/22	No Class - Schedule 1-1 project meetings with instructor
15	2/27	No Class - Schedule 1-1 project meetings with instructor
16	2/29	Technical topic 5: (MAS) Mechamism design	Algorithmic Game Theory, Ch. 9 & 11
17	3/5	Case study 7: Kidney exchange	[Paper 1] Clearing algorithms for barter exchange markets: Enabling nationwide kidney exchanges [Paper 2] FutureMatch: Combining Human Value Judgments and Machine Learning to Match in Dynamic Environments [Optional Reading] Position-Indexed Formulations for Kidney Exchange
18	3/7	Case study 8: Designing an agricultural market	[Paper 1] Allocation for Social Good: Auditing Mechanisms for Utility Maximization [Paper 2] Designing and Evolving an Electronic Agricultural Marketplace in Uganda
21	3/19	Guest lecture - Woody Zhu
22	3/21	Case study 9: Citizen science	[Paper 1] Deep learning with citizen science data enables estimation of species diversity and composition at continental extents [Paper 2] An Animal Detection Pipeline for Identification
23	3/26	Case study 10: NewsPanda - Media monitoring for conservation	[Paper 1] NewsPanda: Media Monitoring for Timely Conservation Action [Paper 2] Where It Really Matters: Few-Shot Environmental Conservation Media Monitoring for Low-Resource Languages
24	3/28	Technical topic 6: (CI) Causal inference and experiment design	The seven tools of causal inference, with reflections on machine learning Elements of causal inference: foundations and learning algorithms
25	4/2	Technical topic 7: (FATE) Algorithmic fairness	Fairness Definitions Explained Classification with fairness constraints: A meta-algorithm with provable guarantees Data preprocessing techniques for classification without discrimination
26	4/4	Case study 11: Homeless and public health service allocation	[Paper 1] Allocating interventions based on predicted outcomes: A case study on homelessness services [Paper 2] Improved Policy Evaluation for Randomized Trials of Algorithmic Resource Allocation
27	4/9	Case study 12: Algorithmic decisions in child welfare	[Paper 1] A case study of algorithm-assisted decision making in child maltreatment hotline screening decisions [Paper 2] Case study: predictive fairness to reduce misdemeanor recidivism through social service interventions
28	4/11	Case study 13: Remote sensing for poverty mapping	[Paper 1] Using publicly available satellite imagery and deep learning to understand economic well-being in Africa [Paper 2] Efficient Poverty Mapping from High Resolution Remote Sensing Images
29	4/16	Guest lecture - Fei Fang: Human Behavior Modeling
30	4/18	Course project presentation

Grading

Course Component	Percentage of Final Grade
Class participation	10%
Case study presentations	20%
Case study writing assignments	20%
Project proposal	5%
Project progress report	10%
Project oral presentation	5%
Project final report	30%

Course Policies

Grading

Late-submission policy: You have a total of 5 late days for case study writing assignments. Late days cannot be used towards project-related assignments. The instructor reserves the right not to grade late submissions beyond the allowed late days.
Re-grading policy: To request a re-grade, please write an email to the instructor titled “Re-grade request from [Student's Full Name]” within one week of receiving the graded assignment.

Collaboration

For case study writing assignments, you may discuss the paper with other students, but you need to specify the names of the students you discussed with in the submission, and complete the writing on your own.
For the course project, you may collaborate with others outside the class (including students, faculty members, and domain experts) with approval from the instructor. If you work in a team of two, or if you have external collaborators, you need to give proper credits to all parties involved, and report the contributions of each contributor in the progress report, final report, and presentations, which will be considered in the grading.

Academic Integrity

For the case study presentations, it is allowed to borrow presentation materials from elsewhere so long as the original source is acknowledged.
For the course project, it is allowed to use publicly available code packages so long as the source of the code package is acknowledged.
Plagiarism in any submitted assignments is strictly prohibited. All writing needs to come from you and do not submit content generated by GPT or similar language models. When in doubt about what you can or cannot use, ask the instructor.
Again, do not plagiarize. There are serious consequences. Please consult SCI's Academic Integrity Country and Pitt's Academic Integrity Guidelines.

Accommodations for Students with Disabilities

If you have a disability for which you are or may be requesting an accommodation, you are encouraged to contact both your instructor and Disability Resources and Services (DRS), 140 William Pitt Union, (412) 648-7890, drsrecep@pitt.edu, (412) 228-5347 for P3 ASL users, as early as possible in the term. DRS will verify your disability and determine reasonable accommodations for this course.

Accommodations for Students with Medical Conditions

If you have a medical condition which will prevent you from doing a certain assignment, you must inform the instructor of this before the deadline. You must then submit documentation of your condition within a week of the assignment deadline.

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. Pitt services are available, and participation in services does work. You can learn more about confidential mental health services available on campus here. Support is always available (24/7) from University Counseling Center: 412-648-7930.

Part of the content adapted from CMU Spring 2023 17-737 AI Methods for Social Good taught by Fei Fang and Fall 2023 94-889 ML for Public Policy taught by Rayid Ghani.