Momentum Centering and Asynchronous Update for Adaptive Gradient Methods

Abstract

We propose ACProp (Asynchronous-centering-Prop), an adaptive optimizer which combines centering of second momentum and asynchronous update (e.g. for $t$-th update, denominator uses information up to step $t-1$, while numerator uses gradient at $t$-th step). ACProp has both strong theoretical properties and empirical performance. With the example by Reddi et al. (2018), we show that asynchronous optimizers (e.g. AdaShift, ACProp) have weaker convergence condition than synchronous optimizers (e.g. Adam, RMSProp, AdaBelief); within asynchronous optimizers, we show that centering of second momentum further weakens the convergence condition. We demonstrate that ACProp has a convergence rate of $O(\frac{1}{\sqrt{T}})$ for the stochastic non-convex case, which matches the oracle rate and outperforms the $O(\frac{logT}{\sqrt{T}})$ rate of RMSProp and Adam. We validate ACProp in extensive empirical studies: ACProp outperforms both SGD and other adaptive optimizers in image classification with CNN, and outperforms well-tuned adaptive optimizers in the training of various GAN models, reinforcement learning and transformers. To sum up, ACProp has good theoretical properties including weak convergence condition and optimal convergence rate, and strong empirical performance including good generalization like SGD and training stability like Adam. We provide the implementation at https://github.com/juntang-zhuang/ACProp-Optimizer.

References in corpus (14)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Decoupled Weight Decay Regularization
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• ADADELTA: An Adaptive Learning Rate Method
• Playing Atari with Deep Reinforcement Learning
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Spectral Normalization for Generative Adversarial Networks
• Self-Attention Generative Adversarial Networks
• On the Convergence of Adam and Beyond
• The relativistic discriminator: a key element missing from standard GAN
• Optimization for deep learning: theory and algorithms
• Closing the Generalization Gap of Adaptive Gradient Methods in Training Deep Neural Networks
• Nostalgic Adam: Weighting more of the past gradients when designing the adaptive learning rate
• AdaX: Adaptive Gradient Descent with Exponential Long Term Memory