(Image credit: Dollar Photo Club.) 

What’s great about the Internet of Things (IoT) is that it opens up a world of real-time services and risk monitoring for insurers. However, those benefits themselves introduce a range of challenges associated with marshalling the enormous amounts of data produced by proliferating sources. This morning IBM launched a set of tools designed to deal with those challenges, the vendor’s Cognitive Analytics for IoT, which it will offer as a set of Watson APIs, as part of its IoT Foundation platform.

(Related: Brave New Underwriting World: Swiss Re Adopts IBM Watson Cognitive Computing)

Cognitive computing, as the name implies, introduces automated processes analogous to human thinking and decision-making. As IBM’s Greg Knowles, writes in the vendor’s Developer blog, “Cognitive analytics enable a new class of systems that interact naturally with end users and understand the intent of both verbal and textual input.” He continues:

These cognitive systems also help us make decisions through reasoning and advance analytics that assert priorities and probable outcomes, and instead of being programmed with explicit logic, they learn from their interactions with us and the surrounding environment, enabling them to keep pace with the volume, variety, and unpredictability of information generated by IoT. Finally, cognitive systems understand unstructured data, like text, video, and images, and can correlate it with machine data to provide greater insights and help us make better, faster decisions.

Cognitive Analytics for IoT is likely to be an important enabler for the advancement of IoT, according to Mark Breading a partner with insurance-focused research and advisory firm SMA (Boston). “That’s because it is not about the things, or even about the data—it is about extracting meaning from the data and taking action,” he comments. “While some of that will be basic sense and response, much of the future value will be based on more sophisticated, reasoned insights.”

IBM describes the Watson APIs within its IoT Foundation platform as follows, characterizing their benefits as allowing clients to make greater sense of burgeoning IoT data through maching learning and correlation with unstructured textual, video and image data:

Natural Language Processing (NLP) enables users to interact with systems and devices using simple, human language. Natural Language Processing helps solutions understand the intent of human language by correlating with other sources of data to put interactions into the context of specific situations. For example, a technician working on a machine notices an unusual vibration. He can ask the system “What is causing that vibration?”. Using NLP and other sensor data, the system automatically links words to meaning and intent, determines the machine he is referencing, and correlates recent maintenance to identify the most likely source of the vibration and recommend an action to reduce the vibration.

Machine Learning automates data processing and continuously monitors new data and user interactions to rank data and results based on learned priorities. Machine Learning can be applied to any data coming from devices and sensors to automatically understand current conditions, what’s normal, expected trends, properties to monitor, and suggested actions when an issue arises. For example, the IBM IoT Foundation platform can monitor incoming data from a fleet of equipment and learn both normal and abnormal conditions. These conditions are often unique to each piece of equipment and its usage conditions, including environment and production processes. Machine Learning helps understand those differences and configures the system to monitor the unique conditions of each asset.

Video and Image Analytics enables monitoring of unstructured data from video feeds and image snapshots to identify scenes and patterns in video data. This can also be combined with machine data to gain a greater understanding of past events and emerging situations. For example, video analytics monitoring security cameras note the presence of a forklift infringing on a restricted area which creates a minor alert in the system. Three days later, the asset in the restricted area begins to exhibit decreased performance. The two incidents can now be correlated to identify a collision between the forklift and asset that was not readily apparent from the video or the data from the machine.

Text Analytics enables mining of unstructured textual data including transcripts from customer calls at call center, maintenance technician logs, blog comments, and tweets to find correlations and patterns in these vast amounts of data. For example, phrases such as “my brakes make a noise”, ”my car seems to slow to stop,” and “the pedal feels mushy,” reported through unstructured channels can be linked and correlated to identify potential brake issues in a particular make and model of car.