Close wants to make machine learning available to all

As companies collect more data from sensors and other sources, it would be nice to be able to analyze it all in real time and take action immediately. Business analysts can analyze, but their discoveries might not happen instantly. This sort of thing is being automated with machine learning, and major web players such as Google (s goog) and Netflix (s nflx) have made advancements in this area. Now startup is looking to democratize machine learning for enterprises and other startups.

Based in Berkeley, Calif., the 1-year-old company just emerged from the enterprise-focused Alchemist Accelerator, and has now entered the Citrix (s ctxs) Startup Accelerator, said Joshua Bloom (pictured), a co-founder of the company and its CEO.

The need for’s service became apparent in the stars. Bloom, who teaches astronomy at the University of California, Berkeley, was working with several other researchers on finding supernovas with telescopes.

It was not uncommon to throw scores of graduate students at the time-consuming task of comparing digital images from telescopes and looking for changes that could point to the occurrence of supernovas, said co-founder Dan Starr. But as telescopes get larger apertures, data volume will increase, and the grad student-powered approach might not scale. So Starr, Bloom and other got students to add their classifications into a system, and they built a machine-learning model that could automatically detect anomalies.

Companies started asking the researchers about using the new machine-learning technique on their own data sets. That’s what led to the establishment of the company, said Starr, the company’s director of data science.

The founders believe software that can learn from training data quickly and get insights at production scale with incoming data in real time would be of value in many industries. Use cases include predictive analytics, real-time fraud detection, automated financial trading and sentiment analysis. The software can run on premise or in the cloud. gets revenue from beta customers. A company working on industrial safety has been using the software to analyze terabytes of sensor data. “What used to take 300 hours of people analysis takes 20 minutes with, at higher accuracy,” Bloom said during his presentation at the Alchemist Accelerator demo day in Santa Clara, Calif., on Thursday.

To be sure, other startups are keen on introducing machine learning to lots of businesses. More use cases will beget more adoption, and the Wise guys are working on that.