Teaching

Advanced technology for Neuromorphic Systems

Graduate course at ETH, ETH, 2024

This lecture traced the evolution of transistor technology from planar CMOS to cutting-edge nanosheet architectures, emphasizing their implications for neuromorphic computing. It analyzed short-channel effects and how FinFETs and FDSOI help mitigate them. The talk then explored the memory technology landscape, from traditional SRAM/DRAM to emerging RRAM, MRAM, and PCM. Special focus was given to In-Memory Computing (IMC), especially its synergy with neuromorphic systems. Finally, the challenges and solutions around variability in RRAM-based spiking neural networks were discussed.

Introduction to Neuromorphic Engineerings

Graduate course at ETH, ETH, 2023

This lecture introduced the core principles of neuromorphic engineering, focusing on the design of real-time, adaptive, and energy-efficient systems that emulate the brain. It explored how sparsity, locality, and event-based computation underpin biological efficiency and how these are translated into hardware. The talk presented neuromorphic sensors, analog and digital implementations, and in-memory computing as enabling technologies. Special emphasis was placed on training and architecture strategies for event-driven systems, including the Mosaic framework and delay-based SNNs. Finally, it highlighted how physical dynamics and enriched computational units can enhance neuromorphic processing.