MIT-SenseTime Alliance Funds Projects to Advance Research In Human and Machine Intelligence
2018-08-28 11:08

Faculty and senior researchers at MIT are teaming up in unprecedented ways to help define the next frontier in human and machine intelligence with projects that delve into everything from fundamental research to societal applications for new technologies.

The MIT-SenseTime Alliance on Artificial Intelligence, which is a program within the MIT Quest for Intelligence, has announced funding for 27 projects involving about 50 principal investigators from departments and labs within engineering, science, architecture, management, and the humanities.

SenseTime, a leading artificial intelligence company founded by MIT alumnus Xiao’ou Tang PhD ’96, jointly created the Alliance with MIT earlier this year to define the next frontier of human and machine intelligence. The selected projects are of one year in duration that are intended to kick-start new efforts and initiate longer-term work. 

 “We were thrilled with the range and creativity of the proposals we received,” says Anantha P. Chandrakasan, dean of the MIT School of Engineering and Vannevar Bush Professor of Electrical Engineering and Computer Science, who is heading the MIT-SenseTime Alliance. 

 “It is particularly exciting to see faculty from so many disciplines join together to embark on projects that speak to the major objectives of the MIT Quest for Intelligence,” he says.

Chandrakasan points to the funded exploratory “moonshot” projects that align with the work of “The Core,” which seeks to advance the science and engineering of human and machine intelligence, explore human intelligence using insights from computer science, and develop new machine-learning algorithms.

One such project features MIT faculty from the Department of Linguistics and Philosophy, the Department of Brain and Cognitive Sciences, and MIT’s Computer Science and Artificial Intelligence Laboratory. These professors are working on a “language moonshot” that explores how insights from linguistic theory can be transformed into machine-learning algorithms to better approximate how people converse. 

Several MIT neuroscientists from the McGovern Institute for Brain Research are joining with an MIT physics professor and others in a moonshot project that explores the biological mechanisms subtending learning and how to incorporate that knowledge into more robust machine-learning techniques. Meanwhile, MIT faculty from computer science and brain and cognitive sciences are collaborating in a project focused on how artificial systems, like robots, can learn common sense knowledge. 

Also popular were projects described as “mission-driven” that aligned, as Chandrakasan notes, with the MIT Quest for Intelligence’s second key entity, “The Bridge,” which is dedicated to the application of MIT discoveries in natural and artificial intelligence to all disciplines.

Seven principal investigators from disciplines including CSAIL, the Department of Aeronautics and Astronautics, and the Department of Earth, Atmospheric, and Planetary Sciences are collaborating on a mission-driven project that will explore drone intelligence for societal applications. 

Another mission-driven project features a faculty member from the Sloan School of Management teaming up with an MIT professor of electrical engineering and computer science who proposes to use machine learning to amplify human intelligence and human employment in future manufacturing. Another involves two MIT mechanical engineering faculty focusing on developing tools in Product Design and Systems Architecture that capitalize on strategies that combine human intelligence with machine intelligence. 

Eight collaborative projects that involve two or more principal investigators also received funding. For example, Media Lab professors will evaluate the use of a social robot as a personalized emotional wellness coach; electrical engineering and computer science professors will use embedded AI to track neurocognitive decline in the elderly; nuclear science and engineering, materials science and engineering, and electrical engineering professors will develop an all-solid device for low-power, fast, brain-like computing; Sloan professors will research dynamic portfolio management with deep learning; and mechanical engineering professors will investigate a specific approach to low-power image regression at the edge for the Internet of Things.

Additional funding was dedicated to 13 projects led by individual faculty from a range of departments, including media arts and sciences, chemistry, computer science, anthropology, physics, brain and cognitive sciences, and more. These projects are wide-ranging and include pursuits such as harnessing artificial intelligence to design vaccines for treating metastatic skin cancer, quantifying the robustness of neural networks, and using AI to increase equity in decision making for agriculture market design. 

“These projects are just the beginning of what promises to be one of the most important initiatives MIT has ever started,” says Antonio Torralba, the director of the MIT Quest for Intelligence. “SenseTime’s tremendous support for pivotal research will deepen the footprint of The Quest on campus.”