SmoothOut: Smoothing Out Sharp Minima to Improve Generalization in Deep Learning

Abstract

In Deep Learning, Stochastic Gradient Descent (SGD) is usually selected as a training method because of its efficiency; however, recently, a problem in SGD gains research interest: sharp minima in Deep Neural Networks (DNNs) have poor generalization; especially, large-batch SGD tends to converge to sharp minima. It becomes an open question whether escaping sharp minima can improve the generalization. To answer this question, we propose SmoothOut framework to smooth out sharp minima in DNNs and thereby improve generalization. In a nutshell, SmoothOut perturbs multiple copies of the DNN by noise injection and averages these copies. Injecting noises to SGD is widely used in the literature, but SmoothOut differs in lots of ways: (1) a de-noising process is applied before parameter updating; (2) noise strength is adapted to filter norm; (3) an alternative interpretation on the advantage of noise injection, from the perspective of sharpness and generalization; (4) usage of uniform noise instead of Gaussian noise. We prove that SmoothOut can eliminate sharp minima. Training multiple DNN copies is inefficient, we further propose an unbiased stochastic SmoothOut which only introduces the overhead of noise injecting and de-noising per batch. An adaptive variant of SmoothOut, AdaSmoothOut, is also proposed to improve generalization. In a variety of experiments, SmoothOut and AdaSmoothOut consistently improve generalization in both small-batch and large-batch training on the top of state-of-the-art solutions.

References in corpus (4)

• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• Don't Decay the Learning Rate, Increase the Batch Size
• Large Batch Training of Convolutional Networks
• Train longer, generalize better: closing the generalization gap in large batch training of neural networks

Cited by in corpus (10)

• SALR: Sharpness-aware Learning Rate Scheduler for Improved Generalization
• Artificial Neural Variability for Deep Learning: On Overfitting, Noise Memorization, and Catastrophic Forgetting
• Study on the Large Batch Size Training of Neural Networks Based on the Second Order Gradient
• S-SGD: Symmetrical Stochastic Gradient Descent with Weight Noise Injection for Reaching Flat Minima
• Taming GANs with Lookahead-Minmax
• Activated Gradients for Deep Neural Networks
• DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing
• Deforming the Loss Surface to Affect the Behaviour of the Optimizer
• BN-invariant sharpness regularizes the training model to better generalization
• Deforming the Loss Surface