View this email in your browser
Looking Back on Our First Year
The MIT Quest for Intelligence launched last spring with a two-part mission: to advance research in human and machine intelligence, and to make AI accessible to all. This past year saw progress on both fronts. With our affiliates, MIT Quest researchers presented results at top AI conferences, and we laid the foundation for integrating AI into more labs and classrooms. We are excited to build on this momentum in the year ahead, and we thank you for your support!


Mentoring Undergraduate Researchers

Our faculty worked with undergraduates on a range of AI-related projects through MIT's Undergraduate Research Opportunities Program (UROP). Meet some of the students who worked on fall UROP and spring UROP projects, which included preparing the home-robot Jibo for a second career as a wellness coach and building an interface for students to train AI models in the cloud.

Creating Accessible AI

We joined forces this year with the student-led MIT Machine Intelligence Community (MIC) to advance our common cause of making AI tools accessible to all. MIC hosts regular AI talks and tutorials for undergraduates, and its members are working with Quest staff and faculty on several projects to expand and promote AI computing on campus. Read a profile of MIC here.


Toward Robust and Interpretable AI

Our fall workshop focused on one of AI's top challenges: building robust and interpretable deep learning models. The promise of autonomous systems like self-driving cars depends on being able to prove that such systems can be made safe and secure. In this Q&A we spoke with workshop organizer and MIT professor Aleksander Madry about new developments in the field.

Exploring the Art and Science of GANs

Our spring GANocracy workshop focused on the promise and opportunities of Generative Adversarial Networks, or GANs, and featured tutorials, talks and posters. Check out a recap of the day in tweets and photos, and read this Q&A with workshop organizer and MIT professor Phillip Isola to understand where GANs are headed. 
Developing Hardware for AI 

Modern AI depends on ever faster computers, but unlike hardware innovations of the past, this next revolution will require a coordinated, multi-disciplinary attack. A workshop organized by MIT professor Jesus del Alamo explored several promising directions, with posters, faculty talks and a panel discussion.


Can Language Models Learn Grammar?

A team of researchers led by MIT professor Roger Levy and MIT-IBM Watson AI Lab researcher Miguel Ballesteros put deep learning models through a set of psychology tests to see what they know about grammar. Answer: more than you might think. Read coverage of the work in MIT News and Venture Beat.

Teaching Machines Visual Reasoning 

A team of researchers led by MIT professor Josh Tenenbaum combine statistical neural networks and symbolic AI to speed up learning and improve transparency in a set of visual comprehension tasks. Read coverage of the work in MIT News and MIT Tech Review.


“I'm an ethicist, and I'm especially interested in these questions around ethics of things we make." National Public Radio features a new MIT course, Ethics of Technology, taught by postdoc Abby Everett Jaques.
“The era of moving fast and breaking everything is coming to a close.” The New York Times features a three-day gathering of AI experts and policymakers at MIT. (Photo: MIT Internet Policy Research Institute).


Explaining the Visual Brain  
Workshop and Challenge: July 19-20, 2019

The Algonauts Project brings together neuroscientists and computer scientists on a common platform to advance both fields. Our two-day workshop will feature talks, posters and tutorials on modeling human and computer vision. Register here.
Copyright © 2019 MIT Quest for Intelligence.
MIT Quest for Intelligence, 400 Main Street, Building E-19
Cambridge, Mass. 02142

unsubscribe | update subscription preferences 

This email was sent to *|EMAIL|*
why did I get this?    unsubscribe from this list    update subscription preferences