Programs that use artificial intelligence to improve forecasting, planning, and decision-making in businesses have been increasing in recent years, but in most cases, they do not have detailed, specific information about the organization itself, which limits the usefulness of those tools.
Devavrat Shah, principal investigator at MIT’s Laboratory for Information and Decision Systems (LIDS), faculty member of the Department of Electrical Engineering and Computer Science (EECS), and member of the Institute for Data, Systems, and Society (IDSS), focuses on how to design methods that can handle split-second decisions using limited computing resources.
“In a sense, with a small amount of resources, you have to do a lot of heavy lifting,” he said. As a researcher, “my interest is in the ability to develop methods that can extract information from data at scale as efficiently as possible.”
Andrew (1956) and Erna Viterbi Professor have been teaching at MIT since 2005.
In 2019, he also founded a spinoff company called Ikigai Labs. Ikigai developed the basic model of the table, time series data based on years of research in Shah’s lab, which was patented and licensed by MIT to the company. This model can take business data inputs from various sources, continuously and averaged, to learn as it goes by testing its predictions against actual results.
Shah explains that the system is an extension of the type of image models used, for example, by GPS devices to convert a small amount of data received from satellites into an accurate model of the earth’s surface, or by a communication system like that in a digital watch that communicates at high speed in an energy-efficient way.
“My interest was: How does one design graphical models for standard, tabular data?” you prune.
While most AI models are taught using text and graphics, this program takes tabular data as its input – structured data like the standard row-and-column format used in spreadsheets. Then it provides a kind of real-time planning, on a much larger scale.
Ikigai’s idea was to provide predictive and decision-making technology to large enterprises, such as consumer goods manufacturers and pharmaceutical companies.
Shah gives an example of how an e-commerce company can use this system.
“Let’s say you make headphones and all kinds of different things. And each product you make has a lot of small parts from different parts of the world. And once the device is sold, it needs to be supported and maintained. And you have to come up with new types of product, you have to market, you have to put the price … places, and what happens to demand if I change the price, or if I launch a promotion?”
He adds that all of these processes are interdependent, and at every stage of the process decisions must be made that have an impact over time. He says, “To a certain extent, digitizing these processes and being able to predict and constantly improve is what leads to better business performance.”
Ikigai was recently acquired by the international company Celonis, where Shah is now a senior scientist in addition to his roles at MIT. Ultimately, he hopes the model he developed for Ikigai will help Celonis deliver tools that can be integrated with company data and business processes to provide real-world analytics that can help make predictions, plans, and decisions.
Shah added that Celonis specializes in digitizing and automating operations for more than 1,400 large companies worldwide. Now that these systems are fully digitized, they provide a platform for Ikigai software to take the next step, to study the data from these digitized systems to provide detailed models to allow simulation of various options, to predict the best strategies, and to predict the results of a specific set of decisions.
“Once the digital layer of these processes is in place and this information is in place,” said Shah, “now, on top of it, we can put the Ikigai stack to be able to make decisions on a much larger scale than otherwise.”
While so many companies are working on various aspects of AI, “we’re very focused on the domain part that the rest of the world is ignoring,” which is the area of structured or time-domain data. By starting with such data, he says, it offers a less expensive version of AI.
He says: “A narrow focus goes hand-in-hand with sharp technology, but it’s broad enough that it’s very important.”
Shah adds, “The latest buzzword appropriate in the modern AI press is ‘world model.’ In a sense, this is trying to create a global model of the business process, so to speak. “
