# An Adaptive and Momental Bound Method for Stochastic Learning

arXiv: https://arxiv.org/abs/1910.12249 · HTML: https://mlsanity.com/paper/1910.12249/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Oct 2019.

## Built on by (most cited first, 12 of 12)

- [Conditional Directed Graph Convolution for 3D Human Pose Estimation](https://mlsanity.com/paper/2107.07797/) (2021)
- [HDFormer: High-order Directed Transformer for 3D Human Pose Estimation](https://mlsanity.com/paper/2302.01825/) (2023)
- [Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attacks](https://mlsanity.com/paper/2105.08714/) (2021)
- [Towards Evaluating the Robustness of Neural Networks Learned by Transduction](https://mlsanity.com/paper/2110.14735/) (2021)
- [AdaNorm: Adaptive Gradient Norm Correction based Optimizer for CNNs](https://mlsanity.com/paper/2210.06364/) (2022)
- [A Survey of Historical Learning: Learning Models with Learning History](https://mlsanity.com/paper/2303.12992/) (2023)
- [AdaInject: Injection Based Adaptive Gradient Descent Optimizers for Convolutional Neural Networks](https://mlsanity.com/paper/2109.12504/) (2021)
- [CV 3315 Is All You Need : Semantic Segmentation Competition](https://mlsanity.com/paper/2206.12571/) (2022)
- [Dragon-Alpha&cu32: A Java-based Tensor Computing Framework With its High-Performance CUDA Library](https://mlsanity.com/paper/2305.08819/) (2023)
- [SING: A Plug-and-Play DNN Learning Technique](https://mlsanity.com/paper/2305.15997/) (2023)
- [Hidden Failure Modes of Gradient Modification under Adam in Continual Learning, and Adaptive Decoupled Moment Routing as a Repair](https://mlsanity.com/paper/2604.22407/) (2026)
- [OmniOpt: Taxonomy, Geometry, and Benchmarking of Modern Optimizers](https://mlsanity.com/paper/2607.04033/) (2026)

## Abstract (arXiv)

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