Sandia National Labs Academic Alliance Collaboration Report 2020-2021


Brad Aimone

Brad Aimone is a computational neuroscience modeling researcher, who also helps lead Sandia’s Neural Exploration Research Laboratory (NERL). NERL enables researchers to explore the boundaries of neural computation and provides a testbed facility for comparative benchmarking and new architecture exploration through use of a variety of neuromorphic hardware and neural algorithms.

(materials and components), which couple well with Sandia’s applied microelectronics research.” UT Austin’s EJ Lee-Furman was instrumental in helping to organize the bio-inspired computing colloquium with Aimone who discussed the next generation of brain-inspired AI at the first event. Aimone noted that the current neural networks revolution was fed by significant advances in computing power and the abundance of training data. However, the expanding scale of neural network models and costs of data collection/management have increased both the energy and monetary costs associated with AI research. Increasingly, researchers need novel, data-efficient AI solutions that achieve cognitive capabilities while leveraging low-power hardware, such as neuromorphic systems. As Aimone explained, Sandia’s research in brain-inspired algorithms enables current AI methods to be deployed on low- power spiking neuromorphic hardware, and also help reach toward novel cognitive capabilities that have eluded the current generation of AI methods. He also described two Sandia-developed tools for porting algorithms to neuromorphic hardware and illustrated how seemingly non-cognitive algorithms can be represented in a spiking framework. Sandia’s recent progress in formalizing neuroscience knowledge of hippocampus subregions into algorithms, he concluded, are suitable for a brain-like memory formation. The technical exchange resulted in a set of working groups in bio-inspired computing and neuromorphic hardware. Currently, each group is developing white papers to help crystallize ideas for future funding opportunities.


2020-2021 Collaboration Report

Made with FlippingBook Ebook Creator