Numerical influence of ReLU’(0) on backpropagationDownload PDF

Published: 09 Nov 2021, Last Modified: 25 Nov 2024NeurIPS 2021 PosterReaders: Everyone
Keywords: Nonsmooth algorithmic differentiation, numerical precision, neural network training, conservative fields
TL;DR: Does ReLU'(0) impact neural network training? Theory and popular thinking suggest not, but this is not verified in practice, due to numerical precision.
Abstract: In theory, the choice of ReLU(0) in [0, 1] for a neural network has a negligible influence both on backpropagation and training. Yet, in the real world, 32 bits default precision combined with the size of deep learning problems makes it a hyperparameter of training methods. We investigate the importance of the value of ReLU'(0) for several precision levels (16, 32, 64 bits), on various networks (fully connected, VGG, ResNet) and datasets (MNIST, CIFAR10, SVHN, ImageNet). We observe considerable variations of backpropagation outputs which occur around half of the time in 32 bits precision. The effect disappears with double precision, while it is systematic at 16 bits. For vanilla SGD training, the choice ReLU'(0) = 0 seems to be the most efficient. For our experiments on ImageNet the gain in test accuracy over ReLU'(0) = 1 was more than 10 points (two runs). We also evidence that reconditioning approaches as batch-norm or ADAM tend to buffer the influence of ReLU'(0)’s value. Overall, the message we convey is that algorithmic differentiation of nonsmooth problems potentially hides parameters that could be tuned advantageously.
Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.
Supplementary Material: pdf
Code: https://github.com/deel-ai/relu-prime
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 3 code implementations](https://www.catalyzex.com/paper/numerical-influence-of-relu-on/code)
18 Replies

Loading