Prof. Milind Tambe

AI for public Safety (2007-2017) 

Prof. Tambe and team provided the first ever applications of computational game theory for operational security. The first of these deployments was the ARMOR system of game theoretic algorithms for security (e.g., counter terrorism) which started operating at the Los Angeles LAX airport in 2007. This is one of the largest and busiest airports in the United States that serves 80 Million passengers, where ARMOR's security games provided day-to-day randomized schedules for checkpoints and canine patrols for over 10 years. This research was followed by pioneering deployments of security games used by major security agencies, including the US Coast Guard to generate patrols in ports such as New York and Boston since 2011, the US Federal Air Marshals for scheduling air marshals on US air carriers international flights since 2009 and others. This research is credited with more than $100M in savings to US agencies.

 

AI for Public Health (2016 - continues)

At Google Research India, Prof. Tambe and the Google team were the first to deploy restless bandits for public health.   This work includes their collaboration with the NGO ARMMAN  which focuses on maternal and child care in India. A key challenge is women drop out of ARMMAN's healthcare programs despite preventable reasons. Prof. Tambe and team are the first to use restless bandits in these large scale public health settings to show a more than 30% reduction in drop out rate.  The system is actually deployed in the field and in continuous use by ARMMAN. Over 50K mothers have been assisted by this system in deployment.  

 

Prior to this work with ARMMAN, Prof. Tambe and his team also provided the first large scale applications of social network algorithms for public health. Conducted In collaboration with social work researchers, their pioneering large-scale field study focused on spreading HIV prevention information among youth experiencing homelessness (YEH). Empirical results with 750 YEH in in Los Angeles demonstrated that their AI-guided interventions led to statistically and practically significant reductions in HIV risk behaviors compared to traditional methods.

Prof. Tambe's research has also focused on fueling and energizing the "AI for social good" ecosystem around the globe. To that end, he and his team at Google Research have started dozens of "AI for social good" projects by matchmaking AI researchers with non-profits; this includes recent kick-off  of 30 new projects across 17 countries. These new projects build on earlier successes at Google Research discussed in this blog post.

 

 

 

Previous Research on Multiagent Systems (before 2007)

Prior to this focused work on "AI for social good", Prof Tambe's early research provided foundational models of multiagent teamwork. His JAIR 1997 "Towards flexible teamwork" won the AAMAS influential paper award (an award given out 10 years after publication of a paper). Tambe and collaborators also the foundational paper Distributed Constraint Optimization (DCOP) that started that subfield by providing the first comprehensive algorithm with quality guarantees (“ADOPT:Asynchronous distributed constraint optimization with quality gurantees” Artificial Intelligence Journal. 2005). Finally, their earlier paper introducing a novel Distributed POMDP model for teamwork won the AAMAS’02 Best paper. During this period, Prof. Tambe and team also built  pioneering systems such as "Electric Elves" team of office assistants, reported in CNN, USA Today and others; built in early 2000s the office agents brought forth a number of challenges of privacy and agent autonomy, while assisting in day to day office activities.

   Prior to Harvard, Prof. Tambe was the Helen N and Emmett H Jones Professor in Engineering at the University of Southern California (USC). At USC, his  pioneering efforts included establishing the Center for AI in Society (CAIS) at the University of Southern California, one of the first centers that focused on an interdisciplinary partnership of AI and social work. At USC he won the USC associates award for creativity in research, the highest honors the university faculty can bestow on its members for distinguished intellectual and artistic achievements.

Prof. Tambe completed his PhD from the School of Computer Science at Carnegie Mellon University. Prior to that, he completed his undergraduate education at BITS, Pilani, India, which awarded him their Distinguished Alumnus Award.

Prof. Tambe's Teamcore group recently celebrated 25 years of research. So far, 35 PhD students have completed their theses under his supervision; he has also mentored 14 Postdoctoral researchers. Currently Prof. Tambe supervises a team of 12 PhD students and postdocs. Prof. Tambe's former students have repeatedly been best thesis award winners or runners-up at top conferences such as AAMAS.  Alumni from Teamcore group are faculty members at CMU MLD, U Chicago, CMU ISR, U Michigan, U Alberta, many other key departments, as well as at key positions in industry.

 

Prof Tambe also manages a group on "AI for social good" at Google Research.

 

Prof. Tambe's service includes National Academy of Sciences panels, blue ribbon panel to review security at LAX airport and DARPA ISAT Panels. He has also served as general co-chair for AAMAS (2004), and as a member of the board of directors of IFAAMAS, the IJCAI executive committee, the board of trustees of RoboCup Robot Soccer World Cup Federation and the steering board of the Conference on Decision and Game Theory for Security (GameSec); he also served as associate editor, advisory board member and editorial board member of journals including the Journal of AI Research (JAIR), IEEE Intelligent Systems and the Journal of Autonomous Agents and Multiagent Systems (JAAMAS), and as co-editor for JAIR special track on AI and Society.

 

In 2013, based on their research, Prof. Tambe co-founded Avata Intelligence, which was acquired by Procore Technologies.