Numenta Newsletter — September 6, 2016
Summer may be winding down, but excitement has been brewing here at Numenta. We
recently unveiled our newly designed website, numenta.com. Those who have
been to our site in the past will recognize that this was not a minor
reconstruction. We underwent a complete design overhaul, all stemming from an
observation by our co-founder, Jeff Hawkins, on the current state of website
design. He said that he found most websites today frustrating – even our own! He
also observed that he almost always started at Wikipedia when trying to learn
about a new company. From that conversation, our website redesign effort was
You’ll see Wikipedia’s influence in our new design, one that we hope makes it
easier for any visitor to find the information they want. You can create your
own experience: walk through each section in order, or start by expanding the
section of your choice. I encourage you to take a tour of the site and get a
feel for the new experience.
While you’re exploring the new design, you may want to stop by the Numenta
blog. There you’ll find several new posts, including a behind the
scenes look at a filming of an interview with Jeff and Yang Lan, a well-known
Chinese broadcast journalist. Yang Lan
and her team are putting together a documentary series on Artificial
Intelligence, and they’ve been traveling the world to interview thought leaders
in the space. The final piece is expected to be seen by roughly a billion
people, and will be released in English as well as Chinese. We’ll share the
interview as soon as it is available.
Another interesting blog post you’ll find is from Jeff, who
reflects on a brain theory essay
he wrote thirty years ago. The ideas he penned in that 1986 paper formed the
basis for the work we are doing at Numenta today. Some of the details may have
changed, but the approach, the enthusiasm and the mission remain the same. Read
the full post to get Jeff’s perspective and a look back at where it all started.
In other news, we recently announced the winners of the
Numenta Anomaly Benchmark (NAB) Competition.
NAB is a benchmark we created to evaluate anomaly detection techniques for
streaming data. It contains a dataset with real-world streams and labeled
anomalies, as well as a scoring mechanism that rewards early detection. We
decided to run a competition because we wanted to continue to grow the
benchmark. We created two categories for prizes:
- Algorithms – new anomaly detection algorithms and the results of running
them on NAB v1.0.
- Datasets – real-world, time-series data with labeled anomalies.
You can read more here
about the winning entries and how we plan to incorporate them into NAB.
Congratulations to all of these winners:
Datasets Category Winners
- 1st place: Samya Bagchi
- 2nd place: BK Ramesh
Algorithms Category Winners
- 1st place: Mikhail Smirnov
- 2nd place: Felix Andrews
- 3rd place: Vladislav Ishimtsev & Evgeny Burnaev