What does the factory of the future look like? Much has been written about 21st century manufacturing, where high tech tools and enterprise apps interconnect everything from the supplier to the distributor, leveraging Big Data, manufacturing intelligence and manufacturing operations management (MOM) systems in a sophisticated interplay of information. Ultimately, the industry is in pursuit of a way to make machines, processes, and products smarter.
But what does that mean? Smart plants, traditionally, have tapped into analytics to extract information from available data. Computers have come a long way in processing power and the programs designed to serve up answers to requests. But it is still based on the 1950s principles of the Von Neumann computing architecture, a rules-based programmatic approach to structured data.
Today, unstructured data, which has no pre-defined data model, (anything from e-mail messages to Word documents to video and social media chatter) is just as important as the numbers in a corporate database. The trick is to capture the unstructured stuff and make it searchable. There’s Big Data, of course, but to truly be intelligent means a computer can’t be bound by logic-based rules of the past.
Instead, the “smart” factory of the future should be using a computing model that is much more responsive. Something that understands natural language and context, and can learn – like a human. According to IBM, manufacturing will need a cognitive machine like Watson – the ultimate artificial intelligence “thinking machine” that appeared on the Jeopardy game show in 2011 and beat two of the show’s biggest champions.
With the formation of IBM’s Watson Group earlier this year, Big Blue announced its commitment to the commercialization of Watson-based apps. To help fuel its efforts, the new unit was given $100 million to invest in third-party software developers, and is making Watson services available in the cloud. Initially, it caught the attention of the healthcare and pharmaceutical industries that are dealing with information overload, which makes accessing relevant content like finding a needle in a haystack. Some healthcare companies, like WellPoint, an insurance provider, are now developing Watson apps that can serve up recommendations for medical services based on a patient’s condition. In pharma, companies are interested in using Watson to help with drug discovery.
And recently, IBM has been exploring how Watson can help manufacturers. According to a NY Times article, IBM researchers began working with Thiess, an Australian contract mining and infrastructure company. It operates a fleet of equipment worth $3 billion, and is looking to Watson to expand predictive maintenance beyond machinery to cover mine operations as whole, factoring in load weights, speed, even weather, terrain, and economic models of mine operations.
Make no mistake, however, that tapping into cognitive technology is no small task. It will require a team of IBM engineers working hand-in-hand with manufacturing IT groups and industry experts to “feed” Watson the information it will need to make decisions. Much of that information likely resides in industrial control systems, like a supervisory control and data acquisition (SCADA), as well as historians, global manufacturing execution system (MES), even 3D modeling and simulation programs.
In order to succeed, IBM will have to work hand-in-hand with manufacturers and their software vendors, and, in the beginning, be very purposeful about where and how cognitive machines can help the manufacturing industry.
Indeed, it won’t happen overnight, but there’s no doubt that artificial intelligence—whether it’s Watson or something else – has a role in the factory of the future.
Do you agree?
Follow Stephanie Neil on Google+.
