MIT Promotes Industrial Competitiveness with MIMO Educational Program

February 28, 2022 by Stephanie Leonida

The Massachusetts Institute of Technology (MIT) brings its Machine Intelligence for Manufacturing and Operations (MIMO) program to the table with supporting research.

The Massachusetts Institute of Technology (MIT) Machine Intelligence for Manufacturing and Operations (MIMO) program aims to use machine intelligence to improve the manufacturing and operations of companies. The program is run by MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and the MIT Manufacturing Lab.

In this news release, the Managing Director of MTI MIMO, Bruce Lawler, speaks to Control Automation about MIMO and the need for machine learning (ML) education and its more rapid deployment across global industries.


Complex industrial equipment

Machine learning and data research. Image used courtesy of MIT MIMO


Machine Learning

Machine learning (ML) is a subset of artificial intelligence that enables computers to learn from experience and data. It can be used in manufacturing and warehouse settings to help with production line optimization, product quality control, quality assurance, as well as predictive maintenance. For ML to be successful in manufacturing, large amounts of data are required for analysis and processing. For example, if you want your ML algorithm to generate insights about which parts are wearing out on your production line, you need access to data about how long each part has been running on the production line and how much wear is happening on each part.

At present, the manufacturing community appears to lack an educated workforce concerning the ins and outs of ML and how it can increase business productivity. As it stands today, data scientists within the manufacturing industry do not have a great deal of crosstalk or the ability to collaborate easily.

The MIMO Program

MIT’s MIMO research and educational program has been developed to promote industrial competitiveness by speeding up ML deployment and helping individuals (from students to industry professionals) understand ML. To help manufacturing companies achieve ML transformation and digitally enhance operations, the MIMO program relies on three key pillars: Research, Education, and Collaboration. The MIMO program incorporates sponsored Ph.D. research into the infrastructure required to support intelligent manufacturing systems and intelligent supply chains. The program also involves research concerning the determination of key performance indicators (KPIs) for ML projects across several industries.



The MIMO program is run by MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and the MIT Manufacturing Lab. Image used courtesy of MIT MIMO


The MIMO Study

In a joint study carried out by McKinsey & Company and MIT for MIMO, 100 companies (with study participants mainly composed of director-level and some vice-president or VP level individuals) were surveyed concerning nine key performance areas. These key areas (Strategy, Opportunity Focus, Budget, Results, People, Data Execution, Deployment, Partnering, and Governance) were selected using bottom-up analysis, which focuses on the fundamental qualities of a business.

Results of the bottom-up analysis were also validated. Bruce Lawler provided additional information in a recent interview with Control Automation. Lawler explained that the study showed participating companies where they ranked against the rest of the survey population. Illustrating the graphical results where dark blue indicates the leaders (75% quartile), with light blue being below average, those less than 50%.

Lawler went on to explain the visualizations generated in the study results also included personalized results for each participant, including other bar chart colors, indications of being in the average quartile, and of average performance in the nine key areas. The study essentially provides a similarity index against the leaders. This study is published in the Harvard Business Review.

Leading the way

The ultimate purpose of the survey, and for the MIMO program, is to provide insights leading to pathways for companies to improve performance through the use of AI concepts. Clearly, observing industry leaders is a major pillar of identifying positive change potential. From the MIMO study, it can be observed that leaders believe certain job roles including AI managers, ML, and data engineers are deemed to be important for successful business operations and growth.

In summary of these results, Lawler says “What you find is that if you see machine learning, engineering, and AI managers, those do not exist at any company, really, other than leaders. But leaders feel they're important.”


ML job titles

The contrast of ML position distributions between leaders and other performers


The MIT study also showed that leading companies put more of their budget towards ML projects for Manufacturing/Operations, Customer Service, Supply Chain, and Product Design. At present, the majority of the industry is comprised of small and medium-sized enterprises (SMEs) and are not investing like the industry leaders who are forging ahead with ML deployment and thus enjoying the largest KPI improvements. This skills gap in the workforce for ML is likely to be stopping SMEs from progressing and improving overall business performance.


ML budget considerations

Budget allocations towards ML projects between industry leaders and other performers


Furthermore, the MIMO study revealed 64% of leaders are storing over 60% of data in the cloud compared with only 25% of companies in the bottom half. Unlike SMEs, which seem to only exclusively use dashboards for data visualization, leading companies appear to use more sophisticated methods including descriptive analytics, predictive analytics, and prescriptive analytics. Leading companies also partner with start-ups in academia to create new ML solutions, as opposed to companies in the bottom quartile of the study, which partner with existing vendors and consultants.


ML data storage

ML data storage comparison between leaders and other performers



As a final conclusion to the indications of this survey, Lawler believes there is significant progress to be made on the part of the smaller businesses, but the solution is not an easy one. “Leaders are ahead by a wide margin and they're getting farther ahead. While most of industry is small and medium businesses, they're not investing and not keeping up, and it's difficult to do. The skills gap is not going to allow them to do that. This indicates to me, you're getting further behind than you know.“

Programs such as MIT’s MIMO will be invaluable for training and upskilling the next generation of data scientists, AI managers, and ML engineers. In this digital age, this initiative will help support manufacturers and operators in making better decisions about their production lines. Advancement is not about simply identifying the traits of the leaders, but leveraging that knowledge to grow yourself and your company to achieve greater results.