MIT CSAIL logo with text that reads machine learning applications initiative

Bringing new applications, workforce development and technological vision to industry.

Initiative Launch

On September 3, 2020, CSAIL Alliances launched our newest research initiative, MachineLearningApplications@CSAIL. Faculty Director Professor Daniela Rus and founding members—Retail Business Services, a subsidiary of Ahold Delhaize; Arrow Electronics; Cisco; and SAP Innovation Center Network—shared the current state of machine learning and their goals for the initiative.

Overview

MIT’s world renowned Computer Science Artificial Intelligence Lab (CSAIL)'s research initiative MachineLearningApplications@CSAIL focuses on applications of the latest machine learning (ML) technologies, potential solutions to the current challenges limiting the abilities of ML, and professional development that will help prepare a company’s workforce for this digital transformation.

Many companies are unsure of how, where, or if they should leverage ML. Awash in data, they are looking to turn that data into intelligence that drives increasingly efficient processes. The valuable insights and impact across all functions from sales, marketing, and customer engagement to logistics, cost control, fraud detection, security, and more can be transformational.

Organizations who know how to leverage and integrate ML across their business will have a competitive advantage. All industries including retail, food/beverage, travel/tourism, household goods, construction, fashion, agriculture, manufacturing/ packaging, education, pharmaceutical, healthcare, and more will all benefit from the latest ML technologies.

MachineLearningApplications@CSAIL will include additional themes as an opportunity for interested companies to gain valuable insights from CSAIL researchers in industry-specific areas. Current themes include:

Learning Robots | Led by Professor Pulkit Agrawal

Advances in robotics have brought us autonomous vehicles, humanoid helpers, and even robotic surgery. But so much more is still possible. Which is harder—to teach a robot to play chess or to use a screwdriver? Manipulations and sensing still have a long way to go. Creating machines that can automatically and continuously learn about their environment is the goal and this theme within MLA@CSAIL focuses on robot learning to enable the next generation.

Therapeutics | Led by Professor Manolis Kellis

There have been many advances in disease detection and treatment, yet so much more work is needed! Globally, there is a substantial unmet need to diagnose various diseases effectively. The complexity of different diseases and the genetic makeup of the patient population are challenges to developing early diagnosis tools and effective treatment to improve patient health. Machine learning can greatly improve the understanding of diseases plaguing the population such as Alzheimer's disease, heart failure, breast cancer, diabetes, obesity, and more. Understanding of the human genome by computational integration of large-scale functional and comparative genomics datasets provides new discoveries and approaches to improve patient outcomes.

Visual Computing | Led by Professor Fredo Durand and Professor William T. Freeman

The field of computer vision and cognition research has exploded recently due to the advancements in sensors, data sets, computing power of GPUs, machine learning and neurally inspired systems of deep learning. The Visual Computing focus will be to revolutionize visual computing systems (including vision, graphics, imaging) and facilitate the transition of fundamental knowledge to real-world technological solutions. Creating machines that can “see” are essential for a wide array of emerging technologies leveraging machine learning such as autonomous vehicles, robotics, and predictive modeling for medical diagnostics.

 

If you are interested in learning more about these themes, please contact alliances@csail.mit.edu.

The Challenge

The ability of a machine to learn depends largely on the accuracy of its underlying mathematical model. Developing and maintaining these models is challenging on several fronts. How can machine learning be leveraged for additional insights, but with outcome guarantees or provability? How can organizations analyze more complex data sets? How can the results be trusted? How can training models be updated with new data to keep the ML systems operating most efficiently?

Organizations must have a skilled workforce of people who can build the models, implement the applications, maintain the models and address ongoing privacy/security issues. From the need for new skills sets to rethinking roles and organizational structures due to automation, companies must take a holistic approach to implementing a machine learning strategy.

Our Approach

Our approach is comprehensive and will address the vision for the future of machine learning, its applications in business, research where no commercial products are currently available and skill development for your workforce. Member companies will engage with our world-renowned lab and:

  • participate in future focused innovation sessions with researchers
  • explore machine learning technology development and pathways
  • advise and inform research that addresses the current challenges limiting the abilities of ML
  • have access to a variety of professional development programs designed to increase worker readiness for the adoption of machine learning in their business
  • connect to startups developing and deploying the latest technologies working to jump start machine learning innovation.

This exciting new initiative will help leaders navigate, consume, digest and prepare their company for all machine learning has to offer.

Faculty Director

Daniela Rus
 
  Daniela Rus

Participating Researchers

Hal Abelson
 
  Hal Abelson

Boris Katz
 
  Boris Katz

Tamara Broderick 
 
  Tamara Broderick

John Leonard

  John Leonard

Randall Davis
 
  Randall Davis
Ankur Moitra
 
  Ankur Moitra
Fredo Durand

  Fredo Durand
Una-May O'Reilly ML

   Una-May O'Reilly
Manya Ghobadi
 
  Manya Ghobadi
Justin Solomon
 
  Justin Solomon
Polina Golland
 
  Polina Golland
Russell Tedrake
 
  Russell Tedrake

Tommi Jaakkola

  Tommi Jaakkola

    
Digital cloud with numerical data on ends of lines
Join the mailing list