NOTE: Numenta has
announced a strategic partnership with Avik Partners,
please read more about the future of
Grok for IT Analytics.
You will notice some substantial changes in our web site this week. I’d like to
take a moment to explain what’s going on. Late last year we shipped our first
product, Grok for IT Analytics on AWS. Since then,
we’ve released several enhancements for the product, including the recent
version 1.5 with a new web charting feature.
Some customers using our product have told us that they have found problems in
their AWS environments that other methods haven’t caught, while others have
found problems hours earlier than they would have otherwise. These results have
demonstrated that our HTM (Hierarchical Temporal Memory) learning algorithms are
performing as we had hoped. We are automatically learning the patterns in the
streaming data of each individual server, and then finding when the input
differs from that expected, thus detecting an anomaly. Other customers are also
using Grok to analyze non-IT data using our custom metrics feature – with
interesting results. We encourage you to explore this option and would love to
hear about your results.
While we will continue to evolve the Grok for IT Analytics product, we are
equally excited about myriad opportunities to apply this technology to other
domains. As a result, over the past few months we have turned our attention
towards creating demonstration applications that show how to potentially apply
our HTM technology to a variety of other problems.
We have restructured our web site to better communicate the breadth of our
technology. Detailed information on Grok for IT Analytics
is still available, but now if you come to visit Numenta, you first will see
information on Numenta technology, before diving deeper to learn more about Grok
and other applications.
As part of the evolution of our web site, we have created some new material that
we hope you will read:
These whitepapers show how the same core technology that finds anomalies in
Grok for IT also finds anomalies in other, very different, streams of data. In
each case, our HTM learning algorithms are automatically modelling different
types of streaming data, learning the patterns, predicting what should come
next, and then highlighting inputs that differ from those predicted. In other
words, our algorithms are doing what your brain does, day in and day out.
We hope you’ll enjoy our new look and refreshed content and look forward to your