An Adaptive and Momental Bound Method for Stochastic Learning

Abstract

Training deep neural networks requires intricate initialization and careful selection of learning rates. The emergence of stochastic gradient optimization methods that use adaptive learning rates based on squared past gradients, e.g., AdaGrad, AdaDelta, and Adam, eases the job slightly. However, such methods have also been proven problematic in recent studies with their own pitfalls including non-convergence issues and so on. Alternative variants have been proposed for enhancement, such as AMSGrad, AdaShift and AdaBound. In this work, we identify a new problem of adaptive learning rate methods that exhibits at the beginning of learning where Adam produces extremely large learning rates that inhibit the start of learning. We propose the Adaptive and Momental Bound (AdaMod) method to restrict the adaptive learning rates with adaptive and momental upper bounds. The dynamic learning rate bounds are based on the exponential moving averages of the adaptive learning rates themselves, which smooth out unexpected large learning rates and stabilize the training of deep neural networks. Our experiments verify that AdaMod eliminates the extremely large learning rates throughout the training and brings significant improvements especially on complex networks such as DenseNet and Transformer, compared to Adam. Our implementation is available at: https://github.com/lancopku/AdaMod

References in corpus (10)

• ADADELTA: An Adaptive Learning Rate Method
• On the Convergence of Adam and Beyond
• The Marginal Value of Adaptive Gradient Methods in Machine Learning
• Regularizing and Optimizing LSTM Language Models
• Improving Generalization Performance by Switching from Adam to SGD
• Adaptive Gradient Methods with Dynamic Bound of Learning Rate
• Adafactor: Adaptive Learning Rates with Sublinear Memory Cost
• Fixup Initialization: Residual Learning Without Normalization
• A Closer Look at Deep Learning Heuristics: Learning rate restarts, Warmup and Distillation
• Nostalgic Adam: Weighting more of the past gradients when designing the adaptive learning rate

Cited by in corpus (5)

• Descending through a Crowded Valley - Benchmarking Deep Learning Optimizers
• Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attacks
• Deep Residual 3D U-Net for Joint Segmentation and Texture Classification of Nodules in Lung
• Conditional Directed Graph Convolution for 3D Human Pose Estimation
• Towards Evaluating the Robustness of Neural Networks Learned by Transduction