NASA Logo, National Aeronautics and Space Administration

+NASA Home

+Ames Home

Biologically Inspired Machine Intelligence Proposal Selected for Award
Intelligent Systems Division Banner

Biologically Inspired Machine Intelligence Proposal Selected for Award

A biologically inspired machine intelligence proposal submitted to the highly competitive 2014 Center Innovation Fund / Director’s Innovation Fund (CIF/DIF) was selected for award. The award will fund a team of researchers to investigate new neurotechnology based computing architectures and algorithms derived from high-fidelity modeling of the mammalian cerebral cortex, which was made available through work by Jeff Hawkin on Cortical Learning Algorithms (CLA). The scientific/technical objective of the proposed research is to conduct a first assessment of the performance of the CLA and compare it to state-of-the-art methods that have been vetted in academic literature and are used ubiquitously in the machine learning community. The team plans to test the capability of the CLA to provide better regression and adverse event-prediction performance than these state-of-the-art machine learning algorithms. CIF/DIF has an acceptance rate of 20%.

BACKGROUND: Jeff Hawkin’s work on the Cortical Learning Algorithms (CLA) is perhaps one of the leading efforts in the field of neurobiologically inspired computing. His work posits that the neocortex is a memory system based on Sparse Distributed Representations (SDRs), learning a sensory-motor model of the world to make predictions and generate behaviors for interacting within its environment. While efforts to implement the CLA in hardware are currently being explored, a detailed implementation of the CLA can now be accessed through the recently open-sourced Numenta Platform for Intelligent Computing (NuPIC).

The proposed research effort offers a unique opportunity for Ames Research Center to be one the first research institutions to assess the CLA since it was open sourced in the summer of 2013. If the overall hypothesis is confirmed, the potential benefit to NASA could fall in a variety of domains, including, but not limited to: autonomous control and navigation for robotic exploration, discovery of anomalous precursors to safety incidents for airborne Unmanned Aerial Vehicles (UAVs) or manually piloted vehicles, and NASA’s high-performance facilities, such as Sustainability Base. This investigation offers a unique opportunity for NASA to have played a pioneering and pivotal role in the “spillover” effect of stimulating deep investigation of this technology in the academic community. If NuPIC does turn out to be a high-performing technology, NASA will be an early adapter, and can thus play a useful role moving forward in further developing applications of this new technology.

NASA PROGRAM FUNDING: ARC Center Innovation Fund / Director’s Innovation Fund (CIF/DIF) program

TEAM: Rodney Martin (PI) and Alexander van Dijk (co-I); R&D staff: Vijay Janakiraman and David Nielsen

COLLABORATORS: Subutai Ahmad (VP Engineering Numenta), Reza Mehrizy (Faculty, Utah Valley University), Nikunj Oza (TI Data Sciences Group Lead), Scott Poll (Sustainability Base Advisor, CS, Code TI), and Steve Zornetzer (Associate Center Director for R&D, CS, Code D)

CUSTOMERS: System-wide Safety Assurance Technologies (SSAT) and Vehicle Systems Safety Technologies (VSST) projects, Aviation Safety Program (AvSP), Aeronautics Research Mission Directorate (ARMD), and the Sustainability Base

Contact: Rodney Martin

First Gov logo
NASA Logo - nasa.gov