On-Chip Error-triggered Learning of Multi-layer Memristive Spiking Neural Networks

Abstract

Recent breakthroughs in neuromorphic computing show that local forms of gradient descent learning are compatible with Spiking Neural Networks (SNNs) and synaptic plasticity. Although SNNs can be scalably implemented using neuromorphic VLSI, an architecture that can learn using gradient-descent in situ is still missing. In this paper, we propose a local, gradient-based, error-triggered learning algorithm with online ternary weight updates. The proposed algorithm enables online training of multi-layer SNNs with memristive neuromorphic hardware showing a small loss in the performance compared with the state of the art. We also propose a hardware architecture based on memristive crossbar arrays to perform the required vector-matrix multiplications. The necessary peripheral circuitry including pre-synaptic, post-synaptic and write circuits required for online training, have been designed in the sub-threshold regime for power saving with a standard 180 nm CMOS process.

arXiv comment15 pages, 11 figures, Journal of Emerging Technology in Circuits and Systems (JETCAS)

References in corpus (8)

• Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing
• SLAYER: Spike Layer Error Reassignment in Time
• Temporal correlation detection using computational phase-change memory
• Training Neural Networks with Local Error Signals
• Mixed-precision deep learning based on computational memory
• A neuromorphic systems approach to in-memory computing with non-ideal memristive devices: From mitigation to exploitation
• Decoupled Neural Interfaces using Synthetic Gradients
• A differential memristive synapse circuit for on-line learning in neuromorphic computing systems