A semantics-preserving co-design exports PyTorch SNNs to event-driven FPGAs via one artifact, achieving 87.4% MNIST accuracy identical to software reference at 0.1375 μs and 31.6 nJ per image.
Fast and energy-efficient neuromorphic deep learning with first-spike times.Nature Machine Intelligence, 3(9):823–835
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Hardware-Software Co-Design for Event-Driven SNN Deployment on Low-Cost Neuromorphic FPGAs
A semantics-preserving co-design exports PyTorch SNNs to event-driven FPGAs via one artifact, achieving 87.4% MNIST accuracy identical to software reference at 0.1375 μs and 31.6 nJ per image.