IBM opens up Watson to analyse feeds from the IoT to drive insights and efficiency
IBM has announced new initiatives designed to bring together the power of its Watson machine learning with the Internet of Things (IoT) in order to help customers and partners, especially those in industry, to reap the potential benefits from insights derived from collecting and analysing data.
IBM said that it is expanding its focus on both cognitive analytics and the IoT with the announcement of a new IBM Watson IoT Foundation Analytics offering that provides developers with application programming interfaces (APIs) to access analytics services delivered from the IBM Cloud.
At the same time, the firm is opening a global headquarters for its new Watson IoT unit in Munich, a facility that will also serve as the first European Watson innovation super centre to drive collaboration between IBM experts and clients. This will be complemented by eight Watson IoT Client Experience Centres spread across Europe, Asia and the Americas.
The move is a continuation of what IBM began earlier this year with the announcement it was investing $3bn into creating an IoT business unit and its work with The Weather Company, but also further back than that to initiatives such as Smarter Cities.
“We’ve been trying to drive more analytics, more data-driven outcomes, and in many ways IoT is a continuation of that,” said IBM IoT vice president Pete Karns. Applying Watson to the problem changes the game, he said.
“If you think of analytics today, you have rules-based analytics, streaming analytics, and predictive analytics, each optimised for a different type of data. With cognitive, you can think of it as a new class of system; it learns at scale, and it consumes multiple types of structured and unstructured data, and as opposed to math-driven analytics, it reasons in a non-biased way to identify things that a human might miss,” he explained.
With this in mind, IBM is providing four new APIs As part of the new Watson IoT Foundation Analytics offering. These comprise the Natural Language Processing API; Machine Learning Watson API; Video and Image Analytics Watson API; and the Text Analytics Watson API.
These can be used independently or in concert with each other to create solutions that can draw on numerous sources of information to deliver recommendations.
“With NLP, say you’re an engineer on a manufacturing line, and you get an alert about a deviation on a piece of equipment, today you would grab a spreadsheet and pull up machine data looking at trends over time, but with NLP, you just ask ‘Watson why is equipment 1234 vibrating out of tolerance?’ And Watson will come back with probabilities and suggestions as to why,” Karns explained.
In the same scenario, Machine Learning would look at the historical data from the sensors and correlate with other information, such as the load on the machine or the environmental conditions, while Video Analytics might pick up that something had collided with the machine recently.
The Text Analytics enables analytics on data such as Twitter feeds, which could be used to warn a vendor of developing potential issues by picking up gripes mentioning its products on social media.
Meanwhile, Karns said that the location of Munich for the Watson IoT unit is because this places it close to many industrial companies that IBM sees as key customers and partners for IoT solutions.
“One of the large domains for IoT is the industrial world. Certainly, there is a tremendous amount of activity in consumer domains, and also in insurance and retail and health, but we’ve got a lot of clients in the industrial sector and when you think about innovation and some of the industrial and auto companies that are here in the Munich area, it’s very logical for us to build on the footprint we already have here,” he explained.
“The way to think about what we’re doing here is to have a campus-like environment, with developers, consultants, researchers, user experience designers, who will be able to interact with clients and partners,” Karns said.
“One thing that is clear about IoT is that we’re all on a journey. What we find is that we can help our clients go faster when we have workshops with the right experts all under one roof,” he added.
IBM’s broader IoT portfolio comprises solutions to help ingest data and manage it at scale, and this is where the acquisition of The Weather Company comes into the picture, according to Karns.
“When you think about what they do, they have a tremendous scale of data collection from weather instrumentation, and they’re serving up anywhere from 10 to 25 billion micro forecasts per day, so their data platform is pretty phenomenal,” he said.
IBM’s new Watson IoT Foundation Analytics services are available immediately from the IBM Cloud.