Convolutional recurrent connections combined with delay learning in SNNs deliver 99% savings in recurrent parameters and 52x faster inference on audio classification with no accuracy loss.
Delgrad: Exact event-based gradients for training delays and weights on spiking neuromorphic hardware,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Combining Convolution and Delay Learning in Recurrent Spiking Neural Networks
Convolutional recurrent connections combined with delay learning in SNNs deliver 99% savings in recurrent parameters and 52x faster inference on audio classification with no accuracy loss.