Adding Gradient Noise Improves Learning for Very Deep NetworksAdding Gradient Noise Improves Learning for Very Deep NetworksArvind Neelakantan and Luke Vilnis and Quoc V. Le and Ilya Sutskever and Lukasz Kaiser and Karol Kurach and James Martens2015
Paper summarydavidstutzNeelakantan et al. study gradient noise for improving neural network training. In particular, they add Gaussian noise to the gradients in each iteration:
$\tilde{\nabla}f = \nabla f + \mathcal{N}(0, \sigma^2)$
where the variance $\sigma^2$ is adapted throughout training as follows:
$\sigma^2 = \frac{\eta}{(1 + t)^\gamma}$
where $\eta$ and $\gamma$ are hyper-parameters and $t$ the current iteration. In experiments, the authors show that gradient noise has the potential to improve accuracy, especially given optimization.
Also find this summary at [davidstutz.de](https://davidstutz.de/category/reading/).
First published: 2015/11/21 (4 years ago) Abstract: Deep feedforward and recurrent networks have achieved impressive results in
many perception and language processing applications. This success is partially
attributed to architectural innovations such as convolutional and long
short-term memory networks. The main motivation for these architectural
innovations is that they capture better domain knowledge, and importantly are
easier to optimize than more basic architectures. Recently, more complex
architectures such as Neural Turing Machines and Memory Networks have been
proposed for tasks including question answering and general computation,
creating a new set of optimization challenges. In this paper, we discuss a
low-overhead and easy-to-implement technique of adding gradient noise which we
find to be surprisingly effective when training these very deep architectures.
The technique not only helps to avoid overfitting, but also can result in lower
training loss. This method alone allows a fully-connected 20-layer deep network
to be trained with standard gradient descent, even starting from a poor
initialization. We see consistent improvements for many complex models,
including a 72% relative reduction in error rate over a carefully-tuned
baseline on a challenging question-answering task, and a doubling of the number
of accurate binary multiplication models learned across 7,000 random restarts.
We encourage further application of this technique to additional complex modern
architectures.
Neelakantan et al. study gradient noise for improving neural network training. In particular, they add Gaussian noise to the gradients in each iteration:
$\tilde{\nabla}f = \nabla f + \mathcal{N}(0, \sigma^2)$
where the variance $\sigma^2$ is adapted throughout training as follows:
$\sigma^2 = \frac{\eta}{(1 + t)^\gamma}$
where $\eta$ and $\gamma$ are hyper-parameters and $t$ the current iteration. In experiments, the authors show that gradient noise has the potential to improve accuracy, especially given optimization.
Also find this summary at [davidstutz.de](https://davidstutz.de/category/reading/).