Reproduce “Stochastic dendrites enable online learning in mixed-signal neuromorphic processing systems” 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
Matteo Cartiglia, Arianna Rubino, Shyam Narayanan, Charlotte Frenkel, Germain Haessig, Giacomo Indiveri, Melika Payvand
Researchers from the University of Zurich and ETH Zurich fabricated a mixed-signal neuromorphic chip incorporating stochastic dendritic online learning circuits. This system achieved over 85% accuracy on the MNIST dataset using 4-bit precision and consumed 455 nW per neuron.
Key insights
- Biologically inspired multi-compartment dendrites provide a plausible and effective architecture for implementing error-based online learning directly in neuromorphic hardware.
- Stochastic rounding is a robust and power-efficient method for translating analog learning signals into discrete, low-precision weight changes compatible with on-chip memory.
- The co-design of algorithms and circuits enables the integration of complex, biologically inspired learning rules into mixed-signal neuromorphic systems for autonomous adaptation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.