Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the revolutionary world of neuromorphic computing in this 16-minute conference talk that demonstrates how spiking neural networks are transforming edge AI applications. Discover how ultra-low-power neuromorphic microcontrollers can process sensor data in real-time using innovative architectures that combine spiking neural networks, compact CNNs, and RISC-V cores—all operating directly on-device to deliver true intelligence where power efficiency, privacy, and minimal latency are critical.
Learn about the technical foundations of encoding IMU data into neural spikes for real-time action recognition, and understand how signal processing pipelines leverage temporal-contrast encoders working alongside integrate-and-fire neurons. Examine the development workflow through a cycle-accurate SoC simulator that enables confident pre-silicon development, and explore the Talamo SDK that provides a PyTorch-like training environment with direct compilation to executable binaries.
Witness a live demonstration of gesture recognition running on wearable devices, showcasing the practical applications of this technology. The presentation concludes with the introduction of a commercial neuromorphic chip specifically optimized for consumer electronics, smart home systems, industrial monitoring applications, and wearable devices, illustrating the real-world potential of bringing brain-inspired computing to edge applications where traditional cloud-based solutions face limitations in power consumption, latency, and privacy requirements.
