Coldlight Solutions (Acquired by PTC)

Internet of Things Machine Learning

Many Internet of Things applications involve massive data sets, far too big for any human to reasonably keep track of, analyze, and interpret. With Neuron, ColdLight Solutions wants to provide sophisticated machine learning and predictive modeling to software at all levels, from cloud services down to individual devices.

Neuron takes in data — from users, device sensors, or web services — via RESTful APIs, and crunches the numbers in various ways. It hunts for interesting signals in the noise, runs simulations, creates and verifies predictive models, and offers insights in the form of “prescriptive intelligence.” That can mean both automatic adjustments to the software providing Neuron’s inputs, or suggestions to human users for how to improve or get more effective use out of the systems Neuron is tracking.

Internet of Things Machine Learning

Machine learning doesn’t remove the human element from data science — it draws on computers’ strengths in handling big data to compliment our understanding of semantics and context. Before deploying Neuron, the team at ColdLight will sit down with a developer and work out exactly what kinds of data the software should pay attention to, and what insights the customer wants to get out of the analysis. Then ColdLight can tailor the application to provide useful results and avoid meaningless correlations.
Internet of Things Machine Learning: Neuron

Neuron can do its processing in the cloud, on local servers, or even on individual devices (assuming they have enough embedded computational power). The last option is attractive for IoT applications, because data stays as local as possible and users get the benefit of direct feedback from whatever Neuron learns.

Learn more about Neuron at or in the video below.

Related: Numenta

Coldlight Solutions (Acquired by PTC)

IoT & Machine Learning, what's next?

Trend research, startups, investment activity