# Adam with Bandit Sampling for Deep Learning

arXiv: https://arxiv.org/abs/2010.12986 · HTML: https://mlsanity.com/paper/2010.12986/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Oct 2020.

## Built on by (most cited first, 11 of 11)

- [1-bit Adam: Communication Efficient Large-Scale Training with Adam's Convergence Speed](https://mlsanity.com/paper/2102.02888/) (2021)
- [The Internet of Federated Things (IoFT): A Vision for the Future and In-depth Survey of Data-driven Approaches for Federated Learning](https://mlsanity.com/paper/2111.05326/) (2021)
- [Survey: Multi-Armed Bandits Meet Large Language Models](https://mlsanity.com/paper/2505.13355/) (2025)
- [KOALA: A Kalman Optimization Algorithm with Loss Adaptivity](https://mlsanity.com/paper/2107.03331/) (2021)
- [Transformer with Memory Replay](https://mlsanity.com/paper/2205.09869/) (2022)
- [Rethinking Teacher-Student Curriculum Learning through the Cooperative Mechanics of Experience](https://mlsanity.com/paper/2404.03084/) (2024)
- [ADAM Optimization with Adaptive Batch Selection](https://mlsanity.com/paper/2512.06795/) (2025)
- [Preferential Subsampling for Stochastic Gradient Langevin Dynamics](https://mlsanity.com/paper/2210.16189/) (2022)
- [When do Convolutional Neural Networks Stop Learning?](https://mlsanity.com/paper/2403.02473/) (2024)
- [Dynamic Learning Rate for Deep Reinforcement Learning: A Bandit Approach](https://mlsanity.com/paper/2410.12598/) (2024)
- [Sample-Adaptive Zeroth-Order Post-Training](https://mlsanity.com/paper/2610.12566/) (2026)

## Abstract (arXiv)

Adam is a widely used optimization method for training deep learning models. It computes individual adaptive learning rates for different parameters. In this paper, we propose a generalization of Adam, called Adambs, that allows us to also adapt to different training examples based on their importance in the model's convergence. To achieve this, we maintain a distribution over all examples, selecting a mini-batch in each iteration by sampling according to this distribution, which we update using a multi-armed bandit algorithm. This ensures that examples that are more beneficial to the model training are sampled with higher probabilities. We theoretically show that Adambs improves the convergence rate of Adam---O(√fraclog nT ) instead of O(√fracnT) in some cases. Experiments on various models and datasets demonstrate Adambs's fast convergence in practice.