CRCS Workshop on AI for Social impact

Date: 

Thu - Fri, Mar 5 to Mar 6, 9:00am - 5:00pm

Location: 

Harvard Faculty Club Reading Room (20 Quincy Street, Cambridge, MA)

 

Recap

CRCS director Milind Tambe welcoming attendees.

Out of over 170 applications, 60 of the brightest rising stars in AI for Social Impact (AISI) were selected to attend the workshop, present their research, network, and be mentored by senior AISI faculty and researchers. Here they are welcomed by CRCS director Professor Milind Tambe.

Dr. Indrani Medhi Thies, Senior Research Scientist at Microsoft Research India, delivered the opening keynote. Dr. Indrani Medhi Thies, Senior Research Scientist at Microsoft Research India, delivered the opening keynote describing her work designing technology for low literate and novice technology users around the globe. She has embedded herself within many communities, especially some in rural India, to design technology that works for everyone.

Kunal Khadilka, CMU gives a lightning talk

   Andrew Perrault, Harvard, gives a conference-style talk

Kunal Khadilka, CMU (above) gives a lightning talk

 Andrew Perrault, Harvard (below), gives a conference-style talk

Talks given by the workshop participants covered a wide breadth of topics such as AI for education, robotics, wildlife, healthcare, and hate speech. All talks were complemented with a poster session on each day, where attendees networked with the presenters and learned more about their work.

a panel on the theme of Creating social impact with interdisciplinary AI researchNeil Gaikwad, MIT (left), moderated a panel on the theme of Creating social impact with interdisciplinary AI research, seated by Harvard’s own Milind Tambe, Cynthia Dwork, and Jonathan Zittrain (left to right). Panelists offered experiences from their distinguished careers in law and computer science about topics such as the importance of AI researchers spending time in the worlds of their interdisciplinary collaborators, how to conduct interdisciplinary work while seeking tenure, and how to effectively manage industry partnerships while avoiding unseen biases.

Prof. Patrick Fowler, Associate Professor in the Brown School of Social Work at Washington University in St. LouisProf. Patrick Fowler, Associate Professor in the Brown School of Social Work at Washington University in St. Louis, gave the second keynote address of the workshop, detailing his work using AI to design interventions to reduce homelessness. Professor Fowler has been championing the use of Bayesian methods that allow policy-makers to ask “what if” questions of historical data, providing valuable insights to inform intervention decisions that reduce the homeless burden in cities.

Distinguished faculty and industry research mentorsDistinguished faculty and industry research mentors dedicated an hour and a half with the rising stars discussing everything from career advice to work-life balance to “the next big thing” in AISI. Left to right: Yiling Chen, Harvard; Pragya Sur, Harvard; Milind Tambe, Harvard; David Parkes, Harvard; Ehi Nosakare, MIT; Sai Ravela, MIT (not pictured).

poster sessionThe rising stars shared and received feedback on their work during a poster session attended by the workshop keynote speakers, Harvard faculty and students. Researchers presented work ranging from socially designed robots to AI for wildlife and even to AI for redirecting angry tweets toward relevant online petitions!

Prof. Matthew Ferrari, Associate Professor of Biology at Penn State UniversityProf. Matthew Ferrari, Associate Professor of Biology at Penn State University, gave the final keynote address. He described his body of research in epidemic modeling which uses AI methods to help policymakers make optimal decisions about how and when to use actions such as treatment, containment, and vaccination.

Nafisat Afolake Adedokun-Shittu, University of Ilorin
Nafisat Afolake Adedokun-Shittu, University of Ilorin, remotely presented from Nigeria her group’s work designing augmented reality systems for education. Their systems are having a positive impact on student retention of material, thanks to the intuitive aspects that AR brings to such subjects as geography and astronomy.

Lily Xu, Harvard (right), moderated a panel seated by Desmond Patton, Columbia; Susan Murphy, Harvard; Erez Yoeli, MIT (right to
Lily Xu, Harvard (right), moderated a panel seated by Desmond Patton, Columbia; Susan Murphy, Harvard; Erez Yoeli, MIT (right to left). Experts shared their diverse viewpoints as: a social work researcher who brings AI tools to identify and help troubled youth online, an AI researcher who likes to prove regret bounds but also embed themself with the struggling patients they try to support, and a behavioral scientist who likes to design systems with an actual incentive for people to generate the data that AI needs to work!

Harvard, MIT, and USC organizers addressed the workshop at its closeHarvard, MIT, and USC organizers addressed the workshop at its close, sealing the inaugural Harvard CRCS Workshop for Rising Stars in AI for Social Impact. Keep an eye out for updates from CRCS about a second edition of this workshop next year!

 

Overview
 

The Harvard Center for Research on Computation and Society (CRCS) Workshop on AI for Social Impact is a two-day event for junior researchers featuring talks and panel discussions from leaders in the fields of Artificial Intelligence, Public Health, and Conservation who are actively engaged in deployed interdisciplinary AI research across the globe. PhD students and Postdoctoral scholars whose research is related to emerging or deployed applications of AI for Social Impact are invited to submit applications to present their work. We also encourage early-stage PhD students to apply who may not have presentations but who are interested in attending. Limited travel funding is available.

Over the course of two days, attendees will have the opportunity to hear from our keynote speakers from cross-cutting domains. Our speakers will also sit on panels with other experts from CRCS, Harvard, and the greater Boston area, engaging in such discussions as “What does it mean to create social impact with AI research?”

Attendees will present their work in one of two formats:

  1. Lightning talk plus poster session.

    We highly encourage attendees to consider these sessions as opportunities to identify new peers and collaborators who share common research or application domain interests.

  2. Conference-length talk.

    Selected senior PhD students and Postdoctoral scholars who submit extended abstracts will be chosen to give conference-length talks on a mature AI for Social Impact project.

Attendees will also participate in one interactive session each day:

  1. Rump session on domain-specific AI for Social Impact.

    Attendees will break out into domain-specific groups to discuss AI for Social Impact within specific application domains of interest. Discussions will be led by facilitators and will culminate with a note-sheet that will be published on the workshop website.

  2. Faculty mentoring session.

    Attendees will be matched in small groups with CRCS faculty to discuss research, the challenges of navigating the emerging field of AI for Social Impact, and more.

The workshop will pay special focus on the development of deployed applications. Attendees will learn from our speakers and panelists about the challenges related to deployment such as building successful interdisciplinary partnerships, translating unexpected challenges of real-world deployment into technical AI problems, and publishing deployed work.

Schedule

Thursday, March 5

09:00 - 09:30 : Breakfast and Welcome

09:30 - 10:15 : Keynote : Indrani Medhi Thies

10:15 - 10:30 : Coffee Break

10:30 - 11:15 : Lightning Talks

11:15 - 12:15 : Long Talks ( Lauren Chambers; Bryan Wilder; Anhong Guo; Sarah Keren )

12:15 - 13:15 : Lunch

13:15 - 14:00 : Keynote : Patrick Fowler

14:00 - 15:00 : Panel : What is Social Impact? (Jonathan Zittrain, Milind Tambe, Cynthia Dwork)

15:00 - 16:30 : Faculty Mentoring Session

16:30 - 17:00 : Lightning Talks

17:00 - 18:15 : Poster Session (in Maxwell-Dworkin Hall Room 119)

18:30 onwards : Dinner (in Maxwell-Dworkin Hall Room 119)

Friday, March 6

09:00 - 09:30 : Breakfast

09:30 - 10:15 : Keynote : Matthew Ferrari

10:15 - 10:30 : Coffee Break

10:30 - 11:30 : Lightning Talks

11:30 - 12:30 : Long Talks (Soheil Eshghi; Andrew Perrault; Shalini Pandey; Elizabeth Bondi; Sam Spaulding)

12:45 - 13:30 : Lunch

13:30 - 14:30 : Panel : Topic TBD (Erez Yoeli, Susan Murphy, Desmond Patton)

14:30 - 15:30 : Poster Session

15:30 - 16:00 : Remote Presentations (Nafisat Afolake Adedokun-Shittu, University of Ilorin; Francis Maduakor, African University of Science and Technology, Abuja)

16:00 - 16:45: Breakout Sessions

16:45 - 17:00 : Closing Remarks

Speakers & Panelists

Keynote Speakers

  • Matthew Ferrari, Associate Professor of Biology at Penn State University
  • Patrick Fowler, Associate Professor of Social Work at the Brown School of Washington University
  • Indrani Medhi Thies, Senior Researcher in the Technology for Emerging Markets Group of Microsoft Research

Panelists

  • Susan Murphy, Professor of Statistics at Harvard University
  • Desmond Patton, Associate Professor of Social Work at Columbia University,
  • Cynthia Dwork, Professor of Computer Science at Harvard University
  • Milind Tambe, Gordon McKay Professor of Computer Science at Harvard University
  • Erez Yoeli, Research Associate at MIT Sloan School of Management
  • Jonathan Zittrain, George Bemis Professor of International Law at Harvard Law School 

Faculty Mentors

Lightning talks & Poster Schedule

Thursday, March 5

Morning Lightning Talk Session (10:30A-11:15A)

  1. Abu Elyounes, Doaa
  2. KhudaBukhsh, Ashique
  3. Jabbari, Shahin
  4. Khadilkar, Kunal A
  5. Ponnapati, Manvitha
  6. Wang, Kai
  7. Huang, Taoan
  8. Keymanesh, Moniba
  9. Lehr, Ted

Afternoon Lightning Talk Session (4:30P-5:00P)

  1. Xu, Lily
  2. Ostrowski, Anastasia K
  3. Killian, Jackson
  4. Garimella, Kiran
  5. Low, Daniel M
  6. Huang, Zijin
  7. Awasthi, Navchetan
  8. Nault, Emilyann
  9. Charpignon, Marie-Laure

Just Poster: Biswas, Roshni

Friday, March 6

Morning Lightning Talk Session (10:30A-11:30A)

  1. Rahmattalabi, Aida
  2. Platt, Eleanor J
  3. Kamwele, Louisa mwikali
  4. Ferber, Aaron M
  5. Gaikwad, Neil
  6. Addlesee, Angus
  7. Basak, Anjon
  8. DiChristofano, Alex
  9. Som, Anirudh
  10. Mate, Aditya
  11. Burli, Pralhad
  12. Sarathy, Jayshree
  13. Hillaire, Garron
  14. Pu, Victoria
  15. Ou, Han-Ching
  16. Macarayan, Erlyn
  17. Aiken, Emily L
  18. Haruna Abdu

Just Poster: Ayman, Afiya and Katherine Hoffmann Pham

Presentation guidelines

We have 60 participants and also expect a number of other students/faculty to drop in. The majority come from a CS background, a number from stats/engineering/math, and a small handful from law/policy/social work/public health. In terms of degree of expertise, ~40 are PhD students, ~10 are undergrad/masters, and ~10 are post-docs/industry researchers.

In your talk, we encourage you to emphasize the social impact (either deployed or potential) of your work and discuss any particular challenges you've experienced when working closely with the domain. Long talk Each talk will be allotted a 15-minute slot, which you should plan as a 12-minute talk and 3 minutes for questions/transition.

We also encourage you to bring a poster to present during the poster sessions. Lightning Talk + Poster: The lightning talk will serve as a quick intro to your work to market yourself and encourage people to learn more at your poster later. Then the poster should be an aid to discuss the AI for Social Impact-style challenges you face in your work. We encourage you to focus on the interdisciplinary/deployment/stakeholder issues you face in your work more so than the detailed technical aspects, but it will also be of interest to other AI researchers what kind of models you are using etc.

Lightning Talk Format

In order to give everyone equal time to present during the lightning talks, the session will be formatted as follows:

  • 3 minutes per presenter
  • Present-in-place (organizer will walk the microphone to you)
  • All session slides will be on organizer’s laptop
  • Organizer will click through your slides

Lightning Talk Requirements

  • 3 minutes max
  • 4 slides max: 1 title slide + 3 content slides
  • Must send slides in PDF format to jkillian@g.harvard.edu by 11:59pm EST March 4.
  • Slides will be compiled on the organizer's laptop the night before the lightning talks, so no need to bring your laptop. In order to give everyone time to present, there will be no exceptions to this. You may not use your own laptop :)

Poster Session Format

The poster session will be formatted as follows:

  • Each lighting talk is accompanied with a poster presentation on the same day. Please see the Schedule section below for the assignments.
  • You will be allocated a space where you present your poster.
  • The organizers also provide poster boards and easels. Posters should be 18" x 24" or 24 x 36". For additional guidance, consult the template provided by CVPR 2019, which also includes advice on designing a good poster.
  • Please print your poster and bring it with you. In case you are not able to print the poster, you may contact one of the organizers ( Neil Gaikwad, gaikwad@mit.edu ) for assistance.

Important dates

  • Submission deadline: Friday, January 10, 2020
  • Decision date: Friday, January 17. 2020
  • Registration deadline: Friday, January 31, 2020
  • Deadline to Submit Lightning Talk Slides: Wednesday March 4, 2020
  • CRCS Workshop on AI for Social Impact: Thursday, March 5 and Friday, March 6, 2020

Objectives
 

  1. Gather junior researchers working on interdisciplinary projects with societal impact to network, foster collaboration, and share ideas and experiences.
  2. Gather experts from Public Health, Conservation and Computer Science to share their impactful work and engage in a dialogue about what it means to make social impact with research.
  3. Create a forum for the discussion of what it means to make social impact with interdisciplinary AI research.

Why attend?
 

Network with like-minded researchers dedicated to social impact. Present your research and get feedback from peers as well as leading experts from AI, public health, and conservation. Share your thoughts and experiences about what it means to have social impact with your research.

Submission Information
 

All submissions must be through the CMT Submission site. Please submit PDFs of (1) your CV, and (2) your application from one of the three options below.

We invite 3 types of applications from PhD students and Postdoctoral scholars:

  1. 1-paragraph abstract describing research related to AI for Society. Topics include, but are not limited to: Public Health, Conservation, Medicine/Mobile Health, Education, and Agriculture. We highly encourage submissions from students in Computer Science as well as related interdisciplinary fields. Accepted submissions will be given a slot for a lightning talk+poster presentation.
  2. 1-page abstract describing a mature AI for Social Impact project, with an emphasis on deployed applications. Topics include, but are not limited to: Public Health, Conservation, Medicine/Mobile Health, Education, and Agriculture. We highly encourage submissions from students in Computer Science as well as related interdisciplinary fields. Six submissions will be selected to give 15-minute talks during the workshop. Preference will be given to senior PhD students and Postdoctoral scholars. Submissions not selected for 15-minute talks will be considered for lightning talks+poster presentations.
  3. 1-paragraph statement describing why you would like to attend the workshop, how it relates to your work so far, and how it will benefit you as a researcher. This is intended for early-stage PhD students who do not have research to present but who are interested in attending the workshop and engaging in discussions. 

All applications from eligible students received by the submission deadline will be considered for travel awards.

We also invite undergraduate students from the local Boston area to apply, though they will not be considered for travel funding. Interested undergraduates should submit either a one-paragraph abstract or statement (in the styles described above).

Submit your application here: https://cmt3.research.microsoft.com/CRCSWAISI2020

Travel awards
 

Limited travel funding is available, under the following stipulations.

  1. Travel expenses will be reimbursed up to $500 per participant.
  2. Accommodations will be provided for out-of-town participants.
  3. We encourage those within a reasonable distance to drive, and carpool if possible.

Organizing Committee
 

Please send any questions or comments to crcs_workshop@g.harvard.edu

Faculty advisory committee