IBM’s Watson is a platform of many talents. It beat Jeopardy! champions Brad Rutter and Ken Jennings in 2011, arguably the biggest victory for Team AI since Deep Blue defeated Gary Kasparov in 1997 and AlphaGo’s triumph over Go expert Lee Seedol last year.
Since its Jeopardy! victory, Watson’s cognitive capabilities have been tapped for a variety of applications, including:
- Cancer Diagnoses
- Product Development
- Teaching Undergraduate Students
- Piloting Autonomous Vehicles
- Making Buildings Smarter
- Improving Your Daily Commute
Now, the system is coming to factory floors with the launch of a new IBM Watson Internet of Things (IoT) solution, Cognitive Visual Inspection. The solution uses Watson to review and analyze parts, components and products, identifying defects by matching patterns to images of defects that it has previously encountered, analyzed and classified.
The company’s goal is to provide manufacturers with a ‘cognitive assistant’ to minimize defects and increase product quality. Early testing suggests it can do just that.
In a test involving a production cycle that typically takes eight days with half a day required for visual inspection, Watson was reportedly able to reduce inspection time by 80 percent and manufacturing defects by 7-10 percent.
Using an ultra-high definition (UHD) camera and cognitive capabilities from IBM Watson, the solution captures images of products as they move through production and assembly. Working with human inspectors, it can detect defects in products, including scratches or pinhole-size punctures.
The solution, which continuously learns based on human assessment of the defect classifications in the images, is a harbinger of the promises of Industry 4.0.
“By bringing cognition to the factory floor, IBM is helping usher in the fourth industrial revolution where entirely new levels of efficiency, flexibility and product excellence in manufacturing can become an everyday reality,” said Harriet Green, general manager, IBM Watson IoT, customer engagement and education.
The new solution was announced at Hannover Messe 2017.
For more information, visit the IBM website.