Talk Abstract
Predictive Analytics with Numenta Machine Intelligence
As sensors integrate with our daily lives, driven largely by the internet of
things (IoT), there is demand for streaming analytics algorithms to provide
insight from this data. Factories, farms, homes, people, and more are being
outfitted with sensors that produce streaming data, but traditional
batch-processing analytics methods don’t suffice. Algorithms must be able to learn and predict online, in real-time. They also must continuously learn and adapt to changing statistics of the environment while simultaneously making predictions.
At Numenta we’ve developed Hierarchical Temporal Memory (HTM), a theory of neocortex implemented in software for machine learning applications. HTM runs online and unsupervised, performing anomaly detection, prediction, and classification on streaming data. HTM can run on wide variety of data streams, from IT server metrics to GPS coordinates. In this talk, Alex will discuss HTM in the context of predictive analytics, presenting real-world use cases.
Schedule
- 6:00pm – Doors open & food/drinks
- 6:50pm – Announcements
- 7:00pm – Talks Start
- 8:30pm – Networking
Alexander Lavin • Research Engineer