Numenta Newsletter March 2017
Numenta Newsletter — March 29, 2017
Last month, our Research team attended Cosyne 2017 (Computational and Systems Neuroscience), to present important new work on our current area of focus within HTM Theory: sensorimotor integration. This research looks at how the neocortex learns through movement. The team delivered a poster titled “3D Object Learning with Cortical Columns,” which outlined how cortical columns learn 3D sensorimotor models of the world. We believe this work is a significant advance in our evolving theory of how the brain works and expect to share more details as it continues to develop. In the meantime, if you’re interested in learning more about it, our Open Source Community Manager, Matt Taylor, recently sat down with Jeff to discuss the latest research in detail, along with a number of HTM topics. You can watch the entire conversation in the video series “HTM Chat with Jeff Hawkins” and join the discussion on the HTM Forum.
Jeff and Subutai delivered a joint keynote at Cornell Silicon Valley 2017 on March 7 titled “Reverse Engineering the Brain for Intelligent Machines.” During their 20 minute talk, they provided an overview of Numenta and covered topics including how our biological AI approach differs from others, the current state of neuroscience research, and what we’ve learned so far. Though the event was for Cornell alumni, the university was kind enough to let us share this video of the keynote.
For those of you familiar with BAMI, our living book on biological and machine intelligence, we recently added two new chapters: Spatial Pooling and Temporal Memory. These chapters offer a collection of resources–from videos to algorithm details and pseudocode–that help explain each concept. Make sure to follow our NumentaBAMI Twitter account if you’d like to be notified about future updates.
In partner news, Francisco Webber, Co-founder of Cortical.io, was featured on the O’Reilly Data Show podcast this month. In this episode, “Natural language analysis using Hierarchical Temporal Memory,” Francisco talked about how he is applying HTM to natural language and big text data problems.
Our partner Grok launched the Applied Machine Learning Exchange designed to provide insights on leveraging machine learning to solve real, human problems. One of the key components is a conversation series, which kicked off March 1 with a discussion on “Machine Learning for the Entire Business Stack” and featured our Director of Engineering, Scott Purdy. You can watch the replay here.