Reproduce “Analog circuits for mixed-signal neuromorphic computing architectures in 28 nm FD-SOI technology” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ning Qiao, Giacomo Indiveri
Researchers at the Institute of Neuroinformatics developed compact and energy-efficient sub-threshold analog synapse and neuron circuits in 28 nm FD-SOI technology. These circuits successfully reproduce both biologically realistic temporal dynamics and fast Rectified Linear Unit (ReLU)-like transfer functions, while effectively mitigating leakage currents for ultra-low-power, event-based neuromorphic computing.
Key insights
- Sub-threshold analog circuits can be effectively designed and operated in 28 nm FD-SOI, overcoming leakage current challenges to achieve ultra-low power consumption.
- Specific circuit techniques are crucial for precise control of pA-nA currents and for realizing long time constants with compact capacitors in scaled technologies.
- A single neuromorphic circuit architecture can be made versatile, emulating both biologically realistic dynamics (millisecond time constants) and high-speed, Rectified Linear Unit (ReLU)-like transfer functions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.