# Noisy Deep Ensemble: Accelerating Deep Ensemble Learning via Noise Injection

arXiv: https://arxiv.org/abs/2504.05677 · HTML: https://mlsanity.com/paper/2504.05677/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 38 references that are in the map.
- First posted on arXiv 8 Apr 2025.

## Builds on (most cited first, 24 of 24)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [SGDR: Stochastic Gradient Descent with Warm Restarts](https://mlsanity.com/paper/1608.03983/) (2016)
- [Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles](https://mlsanity.com/paper/1612.01474/) (2016)
- [Benchmarking Neural Network Robustness to Common Corruptions and Perturbations](https://mlsanity.com/paper/1903.12261/) (2019)
- [Learning without Forgetting](https://mlsanity.com/paper/1606.09282/) (2016)
- [Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets](https://mlsanity.com/paper/2201.02177/) (2022)
- [Generating Sequences With Recurrent Neural Networks](https://mlsanity.com/paper/1308.0850/) (2013)
- [Deep Ensembles: A Loss Landscape Perspective](https://mlsanity.com/paper/1912.02757/) (2019)
- [BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning](https://mlsanity.com/paper/2002.06715/) (2020)
- [Snapshot Ensembles: Train 1, get M for free](https://mlsanity.com/paper/1704.00109/) (2017)
- [Net2Net: Accelerating Learning via Knowledge Transfer](https://mlsanity.com/paper/1511.05641/) (2015)
- [Learning with Pseudo-Ensembles](https://mlsanity.com/paper/1412.4864/) (2014)
- [Why M Heads are Better than One: Training a Diverse Ensemble of Deep Networks](https://mlsanity.com/paper/1511.06314/) (2015)
- [Flipout: Efficient Pseudo-Independent Weight Perturbations on Mini-Batches](https://mlsanity.com/paper/1803.04386/) (2018)
- [Deep Networks Always Grok and Here is Why](https://mlsanity.com/paper/2402.15555/) (2024)
- [NoisyTune: A Little Noise Can Help You Finetune Pretrained Language Models Better](https://mlsanity.com/paper/2202.12024/) (2022)
- [Explicit Regularization in Overparametrized Models via Noise Injection](https://mlsanity.com/paper/2206.04613/) (2022)
- [MotherNets: Rapid Deep Ensemble Learning](https://mlsanity.com/paper/1809.04270/) (2018)
- [Group Ensemble: Learning an Ensemble of ConvNets in a single ConvNet](https://mlsanity.com/paper/2007.00649/) (2020)
- [Weights Augmentation: it has never ever ever ever let her model down](https://mlsanity.com/paper/2405.19590/) (2024)

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

- [Stochastically Perturbed Weights: Ensembles from Deterministic Machine-Learning Weather Models](https://mlsanity.com/paper/2609.08412/) (2026)

## Abstract (arXiv)

Neural network ensembles is a simple yet effective approach for enhancing generalization capabilities. The most common method involves independently training multiple neural networks initialized with different weights and then averaging their predictions during inference. However, this approach increases training time linearly with the number of ensemble members. To address this issue, we propose the novel ``Noisy Deep Ensemble'' method, significantly reducing the training time required for neural network ensembles. In this method, a textitparent model is trained until convergence, and then the weights of the textitparent model are perturbed in various ways to construct multiple textitchild models. This perturbation of the textitparent model weights facilitates the exploration of different local minima while significantly reducing the training time for each ensemble member. We evaluated our method using diverse CNN architectures on CIFAR-10 and CIFAR-100 datasets, surpassing conventional efficient ensemble methods and achieving test accuracy comparable to standard ensembles. Code is available at hrefhttps://github.com/TSTB-dev/NoisyDeepEnsemblehttps://github.com/TSTB-dev/NoisyDeepEnsemble