FutureAnalytics: Cognitive computing in IoT deployments


shutterstock_279586472Cognitive computing is effectively the development of ‘computer as brain’. Cognitive computing systems can handle complex, ambiguous situations, looking for ‘best’ rather than ‘right’ answers. In preparation for sessions at Analytics Experience in Rome, I  caught up with Sue Feldman, co-founder of the Cognitive Computing Consortium,  for a preview.

How does cognitive computing actually work? 

Cognitive computers ‘learn’ from experience. Cognitive computing systems use data, such as that generated by the IoT, and use it to make decisions. But unlike traditional decision-making systems, cognitive systems learn from previous experience, and make better decisions the next time. With more and more data available every day, and new requirements for its extraction, cognitive computing is likely to grow in importance with the IoT.

What about context? 

Cognitive computing allows for context, and not just rules. This means, for example, that it is incredibly useful in fraud detection in the financial sector, because it no longer has to rely on a set of rules. Instead, it can look for patterns, and by learning about customers’ habits, start to forecast likely  (and unlikely) behaviour, improving fraud detection enormously. This ability to predict is also going to be useful in many other areas related to IoT, from improving productivity to better management of maintenance.

Can this help with skills shortages?

Cognitive computing may eliminate some of the data science skills shortage. Companies lack people with the skills to extract and manipulate data. But the rise of cognitive computing means that computers could soon do their own data extraction and even interpretation, communicating with each other at lightning speed, and eliminating the human ‘middle man’. The advantages of this are already starting to emerge in healthcare.

What is the potential impact on IoT?

Cognitive computing frees up people to focus on more important activities. By handling the more routine bits of ‘thinking’ associated with the IoT, cognitive computing is an important element of the IoT. It is not just data scientists whose time will be freed up by cognitive computing, but many others across businesses. Cognitive computing means that line of business managers can be presented with the information they need, in the right form, and at the right time, saving time on decision-making.

Cognitive systems enable more human interactions; in some cases they can act (almost) like people. They can use tone, sentiment and environmental conditions as well as content, to learn to act more appropriately in a given situation. They therefore support deeper and more human interactions between businesses and customers, providing more and more value as the system learns more.

The use of sensors linked to cognitive computing systems has huge potential to improve products and services, and also how businesses operate. From products that can respond to their users, through to improvements to workflow, cognitive computing enables companies to take full advantage of the potential of the IoT.

Overall, cognitive computing seems to be essential to getting the most out of the IoT. The real potential from the IoT may lie in instant action based on sensor data, and that relies on cognitive computing capacity. Collecting and analysing data in real time, and then making decisions about what to do is too complex for traditional systems. Cognitive computing is the only way in which the huge potential of the IoT is likely to be realised.

Can you give us some industry specific examples?

Cognitive manufacturing enables transformation of processes right across the value chain. From design, through the manufacturing process, to supporting customers after purchase, cognitive computing applied to manufacturing offers potential for improvements. Through three main areas—intelligent equipment, cognitive processes and smarter resource optimization—the way manufacturing companies work and serve their customers is changing.

Cognitive computing can also make travel smoother. From improving security, through managing processes such as flying planes, and even making sure that passengers are in the right place at the right time, cognitive computing can help the IoT to improve air travel in particular. The benefits of managing customers are particularly useful, because late passengers are a real problem for security and runway timeslots.

It is expected to improve customer experience across industries. Faster interactions that feel relatively human are likely to be an improvement on much slower responses that are fully human, as far as customers are concerned. Many customer queries are fairly simple, and using cognitive computing to speed up transfer to the right department may be a huge help. With the addition of deeper learning capacity, cognitive computing is likely to take a greater role in improving customer satisfaction.

Learn more

Our recent study of early IoT adoption, several practitioners shared their progress on this journey, and the emerging case for including cognitive tools. We have summarized key challenges and invited our panel of experts to an online discussion on Twitter on Oct 21st, kicking off at 15hrs CET. We hope you can join us. Questions we have put to our experts to kick things off include:

  1. What are the differences between Cognitive Computing, Artificial Intelligence and Machine Learning?
  2. Where are current cognitive investments primarily directed?
  3. What cognitive-driven products/services can we expect by 2020?
  4. How should data scientists prepare to take full advantage of cognitive tools?
  5. What impact will cognitive developments have on the digital gap?

About Author

Adrian Jones

Director, Global Technology Practice

Adrian is Director for Pre-Sales Support in the SAS Global Technology Practice, with a focus on Big Data, High-Performance Analytics and Enterprise Architecture. Adrian looks at how organisations use data and analytics strategically, with a specific focus on optimizing the data and analytics architecture.

Leave A Reply

Back to Top