Joining Forces to Chart the Frontiers of AI

Today, we’re starting a new blog for the MIT-IBM Watson AI Lab, to highlight the work we’re doing at the frontiers of artificial intelligence. The MIT-IBM Watson AI Lab is a unique new partnership between industry and academia, combining the strengths of two institutions with a long history of achievements in AI to advance the field. In this blog, we’ll feature new research publications, technical discussions of emerging work, sneak peeks at technologies in our pipeline, and perspectives on issues surrounding AI’s role in research and society.

A New Model for Engaging with Academia

The MIT-IBM Watson AI Lab is a new model for engagement between industry and academia. Our lab was founded with a $240m commitment by IBM over 10 years to form a truly collaborative industrial-academic lab. Industry has played a significant role in recent advances in AI and machine learning, in part due to the computational resource requirements of modern machine learning, and the commercial potential of AI technology, which is driving investment. At the same time, academia is the undeniable bedrock of AI research, providing a fertile ground for new ideas, and training the next generation of AI scientists. Academia inherently takes the long — and broad — view, seeking advanced knowledge across a range of fields, often with multi-decade time horizons.

Recognizing the complementary strengths of industry and academia, we’ve created a new collaborative model, wherein MIT faculty and IBM researchers jointly conceive and jointly execute projects aimed at advancing fundamental AI technologies. We collaborate with faculty from nearly every department and center within MIT, a testament to the extent to which AI permeates every subfield of science and engineering.

Research Lab as Start-up

IBM Research has long had a presence in Cambridge, and the MIT-IBM Watson AI Lab is swiftly adding to it, with the goal of setting down deep roots in Kendall Square, just steps from MIT. Kendall Square is known for its vibrant startup environment, and in many ways we are running our brand new lab like a startup inside of IBM. We can focus on a small set of objectives; break down large challenges into smaller steps to speed progress; and iterate on promising projects via experimentation. We also have the latitude to work dynamically across disciplines in our research.

Research Pillars

We have organized our collaborative research at the MIT-IBM Watson AI Lab under four pillars: AI Algorithms, AI Applications to Industries, Physics of AI, and AI for Shared Prosperity. Each of these pillars represents a front line in the drive to expand AI technologies so that they reach more parts of our lives, while ensuring they do so in ways that empower users and enrich their lives. In future posts, we’ll provide highlights of work we’re doing under each pillar.

Authored by David D. Cox (L) and Antonio Torralba (R), Directors, MIT-IBM Watson AI Lab