Mechanism Design for Social Good request for proposals

About

The widespread usage of mobile devices and social networks opens unique opportunities to design new mechanisms on top of those networks—mechanisms that lead to meaningful connections which otherwise would not have occurred.

Imagine, for example, a student who attends an economically poor school and wishes to someday become a computer scientist, but lacks direction on how to achieve that goal. Suppose there is a system that connects this student with a college graduate in computer science from the student’s hometown—a graduate who is willing and able to guide this student down a similar career trajectory.

Facebook has numerous products that try to make such positive connections. Some are part of its social good initiative: products that connect mentors to mentees, products that help people connect and support one another with resources and information during crises, products that allow nonprofits and individuals to create charitable giving fundraisers (with the option to match donations to incentivize participation), and blood donation tools that connect donors to those in need. Separately, Facebook Jobs is a product in which job seekers are connected to small businesses, and Facebook Dating is an upcoming product that aims to help people find partners for long-lasting relationships.

Designing the rules of these types of systems falls into the research agenda of mechanism design for social good (MD4SG). Relative to traditional mechanism design, MD4SG (1) restricts itself to mechanism desiderata with some broader benefit to society, and (2) often starts from a domain of practical importance that is relatively unstudied by the mechanism design community. While this call for proposals is distinct from the MD4SG research group and workshop series, we want to provide support for those efforts and further grow that community. To that end, Facebook is pleased to support the MD4SG agenda by offering $150K in research grants in this area.

Researchers should consider the following problem. Suppose there is (1) an existing online platform that is actively used by the population, and (2) an existing set of social ills (e.g., unemployment, disease, poverty, divisiveness, loneliness). How should one design mechanisms on top of such an online platform to build community in a way that alleviates these social ills?

Researchers can submit proposals for any domain in which a successful (privacy-protective) mechanism would have some broader positive impact on society. We describe some example domains below. (Note that the descriptions below are not all fully specified, and at times are highly stylized. We encourage researchers tackling these domains to consider more complex, realistic models when appropriate in their proposals.)

• Mentorship. Suppose one is given a set of people who are all interested in a particular topic or issue, but who self-sort into two groups: those who need support or advice to achieve their goals, and those with the expertise or experience to help. What mechanism for connecting mentors to mentees will result in the greatest overall experience to help form supportive, meaningful relationships?
• Crisis Response. The mission of the Crisis Response team is to increase the survival and speed of recovery for everyone affected by a crisis. One particular feature from this team is Community Help, which tackles the following problem. Suppose a natural disaster has occurred in a community and as a result, people in the community require help and support in the form of information, volunteer services, shelter, food and water, etc.; while other people and organizations in the crisis community want to provide this kind of help and support as efficiently as possible. How should one design a mechanism to allow people to quickly and safely offer and receive help and support?
• Job market matching. Suppose one is given a set of people looking for jobs and a set of small businesses with job openings. What communication protocol and matching mechanism results in a good matching of prospective employees to small businesses?
• Blood donations. Given a set of people in need of blood donations in a broader population, how can one identify possible donors, increase donor participation, and route donors to appropriate health centers to maximize the number of successful blood donations?
• Charitable giving. Suppose one is given a set of charities, a set of people, a set of weights that indicate which people are likely to support which charities, and a fixed budget for matching donations. What donation-matching mechanism will result in the highest total amount donated across charities? How much can outcomes be improved when the mechanism utilizes known information about social ties between people?
• Compassion and civil discourse. Suppose one is given a set of people with diverse political or philosophical beliefs. What systems (consisting of connecting participants, followed by structured conversation) can one design that allow people with non-overlapping beliefs to engage with each other in productive conversation?
• Partnerships and dating markets. Given a set of people who are interested in forming partnerships, and each with unknown preferences over their possible partners, how should one design a mechanism that elicits information from participants and matches them to prospective partners in a way that achieves a desirable outcome?

Aside from proposals that focus on a particular domain, we are also interested in research contributing to broader methodologies that can be applied across multiple MD4SG domains. Research contributions include, but are not limited to, any of the following:

• Tools that allow a practitioner to more easily express their problem in a way that can directly map to a sufficiently compact game representation.
• Methods for eliciting or aggregating preferences about preferred mechanism outcomes.
• Methods for measuring the extent to which desiderata are satisfied from an existing mechanism.
• Advances in automated mechanism design: exploring new objective functions; tackling problems with high dimensionality; efficient computation of outcomes; designing interpretable mechanisms; designing mechanisms with guarantees.
• Methods and tools for problem discovery, e.g., via crowdsourcing or agents that search for market inefficiencies.

Three $50,000 USD gifts will be awarded. Payment will be made to the proposer’s host university as an unrestricted gift.

Proposals should include

• A summary of the project. Provide a maximum 2-page, clear and concise statement explaining the area of focus, a description of techniques, any relevant prior work, and a timeline with milestones and expected outcomes.
• Curriculum Vitae for the principal researchers
• A proposed budget description (1 page) including an approximate cost of the award and explanation of how funds would be spent
• A one-paragraph biography of the principal researchers

Timing and Dates

• Applications are now open. Applications close September 21, 2018, 5:00 pm PST.
• Notification process: Successful awardees will be notified by email.