# WarpAdam: A new Adam optimizer based on Meta-Learning approach

arXiv: https://arxiv.org/abs/2409.04244 · HTML: https://mlsanity.com/paper/2409.04244/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 32 references that are in the map.
- First posted on arXiv 6 Sep 2024.

## Builds on (most cited first, 15 of 15)

- [Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks](https://mlsanity.com/paper/1703.03400/) (2017)
- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [A Comprehensive Survey on Transfer Learning](https://mlsanity.com/paper/1911.02685/) (2019)
- [An overview of gradient descent optimization algorithms](https://mlsanity.com/paper/1609.04747/) (2016)
- [On First-Order Meta-Learning Algorithms](https://mlsanity.com/paper/1803.02999/) (2018)
- [Meta-Learning in Neural Networks: A Survey](https://mlsanity.com/paper/2004.05439/) (2020)
- [On the Convergence of Adam and Beyond](https://mlsanity.com/paper/1904.09237/) (2019)
- [Meta-SGD: Learning to Learn Quickly for Few-Shot Learning](https://mlsanity.com/paper/1707.09835/) (2017)
- [Meta-Transfer Learning for Few-Shot Learning](https://mlsanity.com/paper/1812.02391/) (2018)
- [The Marginal Value of Adaptive Gradient Methods in Machine Learning](https://mlsanity.com/paper/1705.08292/) (2017)
- [Adaptive Gradient Methods with Dynamic Bound of Learning Rate](https://mlsanity.com/paper/1902.09843/) (2019)
- [Meta-Learning with Warped Gradient Descent](https://mlsanity.com/paper/1909.00025/) (2019)
- [Failures of Gradient-Based Deep Learning](https://mlsanity.com/paper/1703.07950/) (2017)
- [Adaptive Gradient Descent without Descent](https://mlsanity.com/paper/1910.09529/) (2019)
- [A Primer on Coordinate Descent Algorithms](https://mlsanity.com/paper/1610.00040/) (2016)

## Built on by (most cited first, 1 of 1)

- [Evolution of Optimization Methods: Algorithms, Scenarios, and Evaluations](https://mlsanity.com/paper/2604.12968/) (2026)

## Abstract (arXiv)

Optimal selection of optimization algorithms is crucial for training deep learning models. The Adam optimizer has gained significant attention due to its efficiency and wide applicability. However, to enhance the adaptability of optimizers across diverse datasets, we propose an innovative optimization strategy by integrating the 'warped gradient descend'concept from Meta Learning into the Adam optimizer. In the conventional Adam optimizer, gradients are utilized to compute estimates of gradient mean and variance, subsequently updating model parameters. Our approach introduces a learnable distortion matrix, denoted as P, which is employed for linearly transforming gradients. This transformation slightly adjusts gradients during each iteration, enabling the optimizer to better adapt to distinct dataset characteristics. By learning an appropriate distortion matrix P, our method aims to adaptively adjust gradient information across different data distributions, thereby enhancing optimization performance. Our research showcases the potential of this novel approach through theoretical insights and empirical evaluations. Experimental results across various tasks and datasets validate the superiority of our optimizer that integrates the 'warped gradient descend' concept in terms of adaptability. Furthermore, we explore effective strategies for training the adaptation matrix P and identify scenarios where this method can yield optimal results. In summary, this study introduces an innovative approach that merges the 'warped gradient descend' concept from Meta Learning with the Adam optimizer. By introducing a learnable distortion matrix P within the optimizer, we aim to enhance the model's generalization capability across diverse data distributions, thus opening up new possibilities in the field of deep learning optimization.