Well into the second decade of the evolution of the Internet of Things, visionary companies are moving beyond the tactical value of being able to add sensors to physical objects and connect them via a network to applications that offer visibility and control.
Two such visionary companies – one a tech giant (Microsoft) and one an ambitious five-year old company based in France (Spinalcom) – will be sharing their work on the cognitive IoT on a panel lead by industry expert Chris Celiberti (CEO at The Infield Group) in a panel at next week’s IoT Evolution Conference and Expo.
The Brains Behind the Operation
As increasing amounts of real-time data are being collected from exponentially more sensors, data scientists and software developers are able to perform more sophisticated and useful analysis that can be used to improve how systems operate, with and without human intervention. The convergence of faster and better processors, more powerful and efficient sensors at the edge, secure and scalable middleware connecting things and local or distributed cloud-based applications, along with the advent of more advance integration through open APIs and SDKs is contributing to stunning creativity in the area of “predictive” systems not just the early control and reporting models.
“The future belongs to the IoT businesses who are not satisfied with the more tactical solutions that marked the first decade of IoT innovation,” Celiberti said. “We’re moving from reporting to prediction, from statistics to machine learning, and from batch processing to real-time systems that can literally improve performance, productivity and business outcomes continuously.”
Jérémie Bellec, the founder and CTO of Spinalcom, will be sharing his company’s approach to the “neuroscience” behind their IoT technologies. “We look at data and analytics as a nerve center that connects in real time the sensors, or actuators, with the brain,” Bellec said. “Our machines and systems are becoming more human; the analog for memory is the database, and the analog for thinking is the analytics.”
Spinalcom’s Spinalcore offering includes a hub that that does the analytic work, provides connectors to analytics, applications, UIs, and devices across nearly every protocol, making it possible to ramp up cognitive IoT solutions in less time. “As humans, we don’t need to stop our system in production to adapt to new experiences and information – and we have developed our technology learning from how humans think, applying it to our reference architecture.”
Straight Out of Microsoft’s R&D Lab
James McCaffrey, Senior Research Software Development Engineer at Microsoft, will be sharing insights from his work on software systems whose designs have been influenced by the behaviors of biological systems, and the massive value that can be created when algorithms, for example, are applied to large scale data mining and analysis initiatives.
McCaffrey will be covering four emerging trends he is following:
- Bio-Inspired Algorithms – combinatorial evolution, simulated protozoa optimization
- Prediction Markets – cyber systems that out-perform all known alternatives for some problems
- Sentiment Analysis – used in augmented sports predictions and financial market predictions
- Deep Neural Networks – responsible for recent breakthroughs in speech and visual recognition
Seeing Through the Fog
The panel will also cover the impact that “AI” combined with IoT will have on compute requirements, at the edge, in transport, and in the cloud. Much is being said and written about “fog” computing and the gains made when processing, particularly real-time processing, is made more efficient.
Companies store and use data with increasing expectations of profitable insight. Without being able to apply the machine learning to solve practical problems and make things work better, data is useless.
As data volumes and concurrency requirements grow, large enterprises are turning to mass scale architectures and may be unnecessarily driving up the cost of cloud, compute, and more.
In order to keep the “brain working” and to support more automation for improvement of performance, and more innovation in triggering human intervention based on policy (for example, sending a physical technician out to fix a problem the IoT closed loop system cannot fix), the development of local and distributed processing solutions – and acceptance of “approximation” vs. requirements for “exact” depending on the application are important to think about now.
Just like the human brain, performance can suffer with “TMI” (too much information) so orchestrating the entire IoT “stack” based on the desired outcomes. Innovations in edge, fog, distributed, and cloud computing will help accelerate all the benefits of cognition inside the next generation of the Internet of smarter things and systems.
McCaffrey, Bellec and Celeberti will be presenting “Distributed Computing and the Cloud” on Wednesday, July 13, from 5 – 5:45 PM as part of the Fog Computing Analytics and Data Track.
Edited by
Ken Briodagh
