Neural and Synaptic Array Transceiver: A Brain-Inspired Computing Framework for Embedded Learning

TL;DR

The paper introduces NSAT, a neuromorphic framework enabling flexible, efficient embedded learning across various algorithms, supporting adaptive autonomous systems with real-time processing capabilities.

Contribution

It presents the NSAT framework that matches algorithmic needs with neural dynamics, facilitating embedded supervised, unsupervised, and reinforcement learning in neuromorphic hardware.

Findings

Supports event-driven deep learning algorithms

Demonstrates tasks like neural simulation and sequence learning

Enables adaptive mobile and robotic systems

Abstract

Embedded, continual learning for autonomous and adaptive behavior is a key application of neuromorphic hardware. However, neuromorphic implementations of embedded learning at large scales that are both flexible and efficient have been hindered by a lack of a suitable algorithmic framework. As a result, the most neuromorphic hardware is trained off-line on large clusters of dedicated processors or GPUs and transferred post hoc to the device. We address this by introducing the neural and synaptic array transceiver (NSAT), a neuromorphic computational framework facilitating flexible and efficient embedded learning by matching algorithmic requirements and neural and synaptic dynamics. NSAT supports event-driven supervised, unsupervised and reinforcement learning algorithms including deep learning. We demonstrate the NSAT in a wide range of tasks, including the simulation of Mihalas-Niebur neuron, dynamic neural fields, event-driven random back-propagation for event-based deep learning, event-based contrastive divergence for unsupervised learning, and voltage-based learning rules for sequence learning. We anticipate that this contribution will establish the foundation for a new generation of devices enabling adaptive mobile systems, wearable devices, and robots with data-driven autonomy.