Layer-wise synapse optimization for implementing neural networks on general neuromorphic architectures

02/20/2018
by   John Mern, et al.
0

Deep artificial neural networks (ANNs) can represent a wide range of complex functions. Implementing ANNs in Von Neumann computing systems, though, incurs a high energy cost due to the bottleneck created between CPU and memory. Implementation on neuromorphic systems may help to reduce energy demand. Conventional ANNs must be converted into equivalent Spiking Neural Networks (SNNs) in order to be deployed on neuromorphic chips. This paper presents a way to perform this translation. We map the ANN weights to SNN synapses layer-by-layer by forming a least-square-error approximation problem at each layer. An optimal set of synapse weights may then be found for a given choice of ANN activation function and SNN neuron. Using an appropriate constrained solver, we can generate SNNs compatible with digital, analog, or hybrid chip architectures. We present an optimal node pruning method to allow SNN layer sizes to be set by the designer. To illustrate this process, we convert three ANNs, including one convolutional network, to SNNs. In all three cases, a simple linear program solver was used. The experiments show that the resulting networks maintain agreement with the original ANN and excellent performance on the evaluation tasks. The networks were also reduced in size with little loss in task performance.

READ FULL TEXT
research
06/16/2021

A Spiking Neural Network for Image Segmentation

We seek to investigate the scalability of neuromorphic computing for com...
research
02/28/2021

Optimal Conversion of Conventional Artificial Neural Networks to Spiking Neural Networks

Spiking neural networks (SNNs) are biology-inspired artificial neural ne...
research
05/10/2021

Analog Neural Computing with Super-resolution Memristor Crossbars

Memristor crossbar arrays are used in a wide range of in-memory and neur...
research
10/13/2021

Efficient Estimation in NPIV Models: A Comparison of Various Neural Networks-Based Estimators

We investigate the computational performance of Artificial Neural Networ...
research
12/09/2022

Reminding Forgetful Organic Neuromorphic Device Networks

Organic neuromorphic device networks can accelerate neural network algor...
research
07/26/2023

Single Channel Speech Enhancement Using U-Net Spiking Neural Networks

Speech enhancement (SE) is crucial for reliable communication devices or...
research
03/15/2017

A Study of Complex Deep Learning Networks on High Performance, Neuromorphic, and Quantum Computers

Current Deep Learning approaches have been very successful using convolu...

Please sign up or login with your details

Forgot password? Click here to reset