Adding energy-efficiency and AI to neuromorphic computing
Spiking neural networks (SNNs) technology has the potential to become a game changer for robotics, AI and IoT. Chips using SNNs meet the industry’s demand for extremely low-power neural networks that are truly self-learning.
Neuromorphic computing has been around since the late 1980s when Carver Mead designed a system containing electronic analogue circuits to mimic neuro-biological architectures present in the nervous system. Artificial neural networks (ANNs) have already proven their worth in a wide range of application domains, like the radar-based autopilot system in the automotive industry. But ANNs consume too much power and decision-making is too slow. Imec at Holst Centre has overcome these drawbacks with SNNs technology.
SNNs operate very similarly to biological neural networks, in which neurons fire electrical pulses sparsely over time, and only when the sensory input changes. As such, energy consumption can significantly be reduced. Chips featuring SNNs emulate the neural structure and operation of the human brain, its probabilistic nature and are capable of dealing with the uncertainty and ambiguity of the natural world – enabling almost instantaneous decision-making. What’s more, the spiking neurons can be connected recurrently – turning the SNN into a dynamic system that learns and remembers temporal patterns. While the chip’s architecture and algorithms can easily be tuned to process a variety of sensor data (including electrocardiogram, speech, sonar, radar and lidar streams), its first use-case will encompass the creation of a low-power, highly intelligent anti-collision radar system for drones that can react far more effectively to approaching objects.
There is great potential for extremely low-power neural networks that truly learn from data and enable personalised AI. This technology could be used in robotics scenarios in the deployment of automatic guided vehicles (AGVs) and even in health monitoring, since SNNs and our nervous system speak the same language. Therefore the development and application of SNNs marks a significant breakthrough in the biggest tech trends of our time: AI, robotics and IoT.