vege - Fotolia
LAS VEGAS -- About 17,000 people attended this week's IBM World of Watson conference. Many of them were trying to figure out what exactly to do with the cognitive computing engine.
The Watson technology wasn't the only focus here. After an initial Watson-specific conference drew just over 1,000 people to Brooklyn, N.Y., in May 2015, IBM this year combined that event into its much larger IBM Insight analytics conference and changed the overall name to World of Watson. But the growing interest shown in Watson by attendees is a testament to how hot all things related to artificial intelligence are right now.
At the same time, many businesses are just starting to think about how they can use cognitive applications like Watson. While several large, multinational corporations, including General Motors, Teva Pharmaceuticals and Staples, described examples of how they plan to implement Watson technologies, the more typical smaller enterprises were less clear.
Conference attendee Daniel Bertha, vice president of the information control department in the Americas division of Tokyo-based Sumitomo Mitsui Banking Corp., said he could envision some benefit for his business down the road. For now, however, he's more focused on the basics of data quality, a necessary step before engaging in cognitive computing or artificial intelligence.
"You have to build the foundation before you can build the house," Bertha said.
In a presentation, Neil Gregory, reliability engineering manager at New Zealand-based Meridian Energy, said he sees a lot of value in the idea of applying cognitive analytics to the large data sets his company maintains, which stretch back 80 years. But for the moment the company is directing more resources toward developing predictive models that identify wind and solar power assets that are likely to require maintenance. He said he believes the company will move toward cognitive applications in the next two or three years, but for now the focus is on more traditional analytics.
For Jeff Wright, vice president and customer journey leader at BMO Financial Group, cognitive computing could become a useful tool, but the focus has to be on concrete customer benefits.
"I don't think you should ever do a cognitive project just to do it," he said in a panel discussion. "You should identify a customer need and build a product around that." He added that Bank of Montreal, a subsidiary of BMO Financial Group, is currently trying to give its customer service systems a more human feel, and he thinks cognitive computing could be a part of that. But for now it's still in the investigational phase.
Notable Watson-related news
While actual implementations of Watson remain in their infancy, IBM is trying to change that. The company unveiled new features and services at the conference designed to encourage adoption.
The biggest updates are the Watson Data Platform and Watson Machine Learning Service. The platform allows users to ingest large data sets at speeds of up to 100 gigabytes per second into the IBM Cloud. There they can combine data sets, perform data preparation tasks, assign user roles for accessing and collaborating on the data and pull the data into analytics workbooks.
The Data Platform is built around an open architecture and can be accessed using SQL, Python, R, Java and Scala.
The Machine Learning Service works in conjunction with the Data Platform and is intended to be a self-service machine learning tool. It uses the Watson cognitive engine to automatically build machine learning models from data sets. Watson recommends machine learning algorithms based on the data being used. Users can choose from a list of algorithms that best fit the job they are trying to accomplish.
Watson's API could help increase the use of cognitive computing
How big data is affecting the growth of cognitive computing
An expert looks at the impact of cognitive computing