The Internet of Things is generating an explosion of data—much more than companies using traditional tools can analyze. Simularity has come up with a solution. By creating an AI that can learn and do anomaly detection and failure analysis, they enable us to see trends and solve problems in real-time. We sat down with the CEO and founder of startup Simularity, Liz Derr, to give us the details of this disruptive technology.
What is Simularity’s founding story?
Early on, we figured out that a scalable correlation engine technology could apply to many industries.
But we came to a challenge: our engine didn’t have a good sense of time. We researched the algorithms that come with traditional machine learning packages and found that they were not as accurate.
So we set out to build something better. We looked into AI using time series data and developed our own algorithms for spatio-temporal reasoning.
What was your “Aha” moment when you decided to use Big Data for event prediction and anomaly detection in the area of the Internet of Things (IoT)?
The real “aha” moment came when we realized how important time series data is to the Internet of Things.
IoT is generating an explosion of time series data going into the cloud that needs to be analyzed. The real solution is to do the predictive analytics as close to the sources of the data as possible.
So rather than sending all that raw data to the data center, you could just send anomaly scores and incident predictions instead.
Artificial Intelligence technology is a hot topic for 2016. Tell us what role AI plays in your technology.
AI is key to what we do. Unlike machine learning, our AI does its learning on its own. This is important for IoT because every device has different sensors and patterns that are predictive of failures or other problems.
In order for the data to have value, it must be analyzed for trends that can indicate problems or opportunities. The answer is to have an AI that can automatically figure out how to analyze any type of data–and find predictive, complex inter-dependent temporal signals that humans just can’t see.
What was it like working and winning with Yamaha for the 2015 Meeting of the Minds hackathon? Any key takeaways from working with one of the leading motor corporations?
Yamaha’s vision of “Vehicle as Probe” really inspired me. My team came up with the winning idea to attach acidity sensors to all watercraft Yamaha makes, so that scientists can study the rate of ocean acidification (a side effect of global warming), and come up with a solution. I’m impressed with Yamaha’s sincere desire to give back and protect the environment.
The idea we had was to use the Yamaha VASP that measures air quality, and attach it to city vehicles to get a hyper-local view of air quality throughout the city, in real time. We found out that there were higher levels of pollution in some wealthy neighborhoods than there were in many poor ones, which was a real surprise to the city managers.
Since we won the hackathon, Chevron has requested our data, and we are working on defining a similar pilot program with another California city.
What are your goals while you are in Cisco EIR?
Our main goals are to get integrated with existing Cisco technology in the Data and Analytics Group. The recognition and publicity we have received from being in the EIR program has been invaluable.
What is your startup’s greatest accomplishment thus far?
A recent pilot project revealed that our technology could save the customer oil and gas company approximately $2M per DAY. Needless to say, they found this very compelling, and we are working on rolling out our technology to monitor many more of their wells to start saving them that money!
What is your #1 advice for other aspiring entrepreneurs?
Talk to everyone. You never know who is going to turn into a customer, or who is going to be able to introduce you to someone who can help your startup grow. My second piece of advice is to be prepared for the lean times to last years. Most startups fail because the founders give up. Manage your business wisely enough so that you can survive until you find the right product/market fit.
Tags: AI, analytics, Artificial Intelligence, Big Data, entrepreneur, startup