As we approach the end of the year, I’m happy to announce the release of two new resources. The first is a peer-reviewed paper published in Frontiers in Computational Neuroscience. The second is an episode of the Brain Science podcast featuring an interview with our co-founder, Jeff Hawkins. While the podcast is also intended for a neuroscience audience and does get into some technical details, we hope you’ll find it’s an enjoyable conversation that is accessible to non-technical listeners as well.
The new paper, The HTM Spatial Pooler – A Neocortical Algorithm for Online Sparse Distributed Coding, describes an important element of HTM, the HTM spatial pooler, which is a key component of our sequence memory and sensorimotor inference models.
Given our focus on brain theory and research, our goal is to create a set of peer-reviewed papers documenting all of our research discoveries. The HTM spatial pooler (SP) paper marks the fifth peer-reviewed paper in that set. It provides analysis that supports our theory of how the cortex learns predictive models of sequences—both extrinsic and sensorimotor.
Specifically, the paper demonstrates how the SP, a neurally inspired learning algorithm for creating sparse distributed representations (SDRs), supports essential brain computations like sequence learning and memory. It details the SP’s computational properties and value in end-to-end HTM systems. We invite you to read the paper and let us know what you think, through our website or the HTM Forum.
The new podcast is the latest episode of Brain Science with Dr. Ginger Campbell, titled “Jeff Hawkins Explores a New Theory of Cortical Function.” Dr. Campbell has shared that her series, which focuses on neuroscience discoveries and features interviews with scientists around the globe, was partially inspired by Jeff and his book On Intelligence. After appearing as one of her first interviewees many years ago, Jeff returned to spend an hour discussing our latest research and the “missing ingredient” that has been completely missed from cortical theory: a location signal.
As Jeff explains to Dr. Campbell, “As soon as you add this location signal, then all kinds of things make sense, and all kinds of mysteries get resolved, and all of a sudden we can understand what all these layers are doing, and it tells us that the cortex, even a single column of the cortex, is much more powerful than people thought.”
Lastly, we are currently accepting applications for our Spring/Summer 2018 Research Internship positions. If you know of outstanding PhD students or research scientists, encourage them to review the requirements and apply through our website.
Thank you for your continued interest in Numenta.