An Ultra-Low-Power Neuromorphic Spiking Neuron Design
From iis-projects
Contents
Description
Integrate-and-Fire (IAF) neurons are widely available in neuromorphic spike-processing applications. Analog LIF neurons are considered a breakthrough to realize fully analog, fully asynchronous spiking neural network implementation for ultra-low-power AI-embedded systems. In this project, you will seek to realize an analog ultra-low-power (thus low energy/spike) LIF neuron, having a very low-frequency spiking rate (e.g., <100Hz). The prerequisite course is Analog integrated circuits. If the design result is promising, the study will be summarized and submitted to IEEE conferences such as ISCAS/BioCAS.
- Looking for master or semester thesis students
- Supervisor: Kwantae Kim <kimkwa@ethz.ch>
Prerequisites
- Analog integrated circuits
Character
- 20% Literature review
- 20% Theory
- 60% Circuit Design
Professor
Prof. Taekwang Jang <tjang@ethz.ch>
Reference
[1] S. Kim, S. Kim, S. Um, S. Kim, K. Kim and H. -J. Yoo, "Neuro-CIM: ADC-Less Neuromorphic Computing-in-Memory Processor With Operation Gating/Stopping and Digital–Analog Networks," in IEEE Journal of Solid-State Circuits, vol. 58, no. 10, pp. 2931-2945, Oct. 2023, doi: 10.1109/JSSC.2023.3273238.
[2] A. Rubino, C. Livanelioglu, N. Qiao, M. Payvand and G. Indiveri, "Ultra-Low-Power FDSOI Neural Circuits for Extreme-Edge Neuromorphic Intelligence," in IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 68, no. 1, pp. 45-56, Jan. 2021, doi: 10.1109/TCSI.2020.3035575.
[3] S. Vuppunuthala and V. S. Pasupureddi, "3.6-pJ/Spike, 30-Hz Silicon Neuron Circuit in 0.5-V, 65 nm CMOS for Spiking Neural Networks," in IEEE Transactions on Circuits and Systems II: Express Briefs, doi: 10.1109/TCSII.2023.3324584.
