What question did this study set out to answer?

The research aims to optimize spiking neural networks by enhancing training methods for weights and delays.

February 23, 2026Open Access

DelGrad: exact event-based gradients for training delays and weights on spiking neuromorphic hardware

View Full Paper

Ask AI

Bookmark

Authors

Analytical training method improves synaptic weights in spiking neural networks on neuromorphic hardware, suggesting better efficiency.

The research aims to optimize spiking neural networks by enhancing training methods for weights and delays.
Proposed DelGrad analytical training method for spiking neural networks
Implemented on BrainScaleS-2 mixed-signal neuromorphic platform
Focused on event-based computation of loss gradients for weights and delays
Demonstrated parameter efficiency and accuracy benefits from introducing delays
Improved on previous results for training spiking neural networks
Showcased stabilizing effects on noisy neuromorphic hardware

Göltz et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

February 23, 2026Open Access

View Full Paper

Ask AI

Bookmark

Authors

Analytical training method improves synaptic weights in spiking neural networks on neuromorphic hardware, suggesting better efficiency.

The research aims to optimize spiking neural networks by enhancing training methods for weights and delays.
Proposed DelGrad analytical training method for spiking neural networks
Implemented on BrainScaleS-2 mixed-signal neuromorphic platform
Focused on event-based computation of loss gradients for weights and delays
Demonstrated parameter efficiency and accuracy benefits from introducing delays
Improved on previous results for training spiking neural networks
Showcased stabilizing effects on noisy neuromorphic hardware

Göltz et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark