Bay Area HTM Meetup

Welcome to the community hosted HTM [Hierarchical Temporal Memory] Meet up

Meet HTM hackers, AI researchers, Entrepreneurs and enthusiast working in the field of Artificial Intelligence. A brief agenda is listed below. Like before we have a special focus on HTM projects and talks.

Make sure you are signing up for a lightning talk if you plan to present. [Slots are limited, make sure you sign up fast].

Agenda:

• 6:30 - 7:00: Welcome and Sign In

• 7:00 - 7:10: About Trustly [A message from our host for the day]

• 7:10 - 7:20: Message from Numenta [Christy Maver]

• 7:20 - 7:30: State of HTM Open Source [Matt Taylor]

• 7:30 - 8:30: Lightning Talks and Demos

• 8:30 - 9:30: Networking

Lightning Talks and Demos

• See HTM from Multiple Angles : HTMs ought to have flight panels. If you're someone who spends your time making sense of what HTMs are doing, it's essential to be able to quickly see the HTM from different angles. This talk will run through some tools and visualizations that attempt to fill this role. Special thanks to Felix Andrews for creating lots of what I'll show.

• Advanced HTM processing with Apache Flink : Talk will present a new integration project - htm.flink - that enables HTM processing of Flink data streams. We'll look at how Flink solves many of the above challenges and simplifies HTM programming. Note that the project is based on htm.java.

• InfluxHTM: The World as a Temporal Data Stream (Matt) : HTM systems can provide real-time analysis when attached to live temporal data streams, but HTM runtimes can be tedious to configure. We need a common interface for streaming data storage, retrieval, and aggregation for HTM systems, allowing them to contribute their analysis results back into the stream itself. InfluxHTM provides a concrete example implementation of that interface for InfluxDB.

• SmartThings and HTM (Jeff and Matt): Displaying SmartThings IoT sensor data and HTM anomaly indications in real-time.

To sign-up for a demo/lighting talk email: chandan.maruthi@gmail.com

What is HTM?

Hierarchical temporal memory (HTM) is an online machine learning model developed by Numenta, Inc. that models some of the structural and algorithmic properties of the neocortex. HTM is a biomimetic model based on the memory-prediction theory of brain function described by Jeff Hawkins in his book On Intelligence. HTM is a method for discovering and inferring the high-level causes of observed input patterns and sequences, thus building an increasingly complex model of the world.

More information about HTMs: http://numenta.org/htm-white-paper.html