# A New Perspective for Understanding Generalization Gap of Deep Neural Networks Trained with Large Batch Sizes

arXiv: https://arxiv.org/abs/2210.12184 · HTML: https://mlsanity.com/paper/2210.12184/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 50 references that are in the map.
- First posted on arXiv 21 Oct 2022.

## Builds on (most cited first, 22 of 22)

- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [Aggregated Residual Transformations for Deep Neural Networks](https://mlsanity.com/paper/1611.05431/) (2016)
- [Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour](https://mlsanity.com/paper/1706.02677/) (2017)
- [On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima](https://mlsanity.com/paper/1609.04836/) (2016)
- [Exact solutions to the nonlinear dynamics of learning in deep linear neural networks](https://mlsanity.com/paper/1312.6120/) (2013)
- [Don't Decay the Learning Rate, Increase the Batch Size](https://mlsanity.com/paper/1711.00489/) (2017)
- [Train longer, generalize better: closing the generalization gap in large batch training of neural networks](https://mlsanity.com/paper/1705.08741/) (2017)
- [Deep Learning without Poor Local Minima](https://mlsanity.com/paper/1605.07110/) (2016)
- [Measuring the Effects of Data Parallelism on Neural Network Training](https://mlsanity.com/paper/1811.03600/) (2018)
- [Revisiting Small Batch Training for Deep Neural Networks](https://mlsanity.com/paper/1804.07612/) (2018)
- [Hessian-based Analysis of Large Batch Training and Robustness to Adversaries](https://mlsanity.com/paper/1802.08241/) (2018)
- [Deep linear neural networks with arbitrary loss: All local minima are global](https://mlsanity.com/paper/1712.01473/) (2017)
- [UMDFaces: An Annotated Face Dataset for Training Deep Networks](https://mlsanity.com/paper/1611.01484/) (2016)
- [Large-Scale Distributed Second-Order Optimization Using Kronecker-Factored Approximate Curvature for Deep Convolutional Neural Networks](https://mlsanity.com/paper/1811.12019/) (2018)
- [Four Things Everyone Should Know to Improve Batch Normalization](https://mlsanity.com/paper/1906.03548/) (2019)
- [Extrapolation for Large-batch Training in Deep Learning](https://mlsanity.com/paper/2006.05720/) (2020)
- [Extremely Large Minibatch SGD: Training ResNet-50 on ImageNet in 15 Minutes](https://mlsanity.com/paper/1711.04325/) (2017)
- [An Empirical Study of Large-Batch Stochastic Gradient Descent with Structured Covariance Noise](https://mlsanity.com/paper/1902.08234/) (2019)
- [The Limit of the Batch Size](https://mlsanity.com/paper/2006.08517/) (2020)
- [The Effect of Network Width on the Performance of Large-batch Training](https://mlsanity.com/paper/1806.03791/) (2018)
- [Inefficiency of K-FAC for Large Batch Size Training](https://mlsanity.com/paper/1903.06237/) (2019)

## Built on by (most cited first, 3 of 3)

- [GMTRouter: Personalized LLM Router over Multi-turn User Interactions](https://mlsanity.com/paper/2511.08590/) (2025)
- [Novel Deep Neural Network Classifier Characterization Metrics with Applications to Dataless Evaluation](https://mlsanity.com/paper/2407.13000/) (2024)
- [Differentiable Zero-One Loss via Hypersimplex Projections](https://mlsanity.com/paper/2602.23336/) (2026)

## Abstract (arXiv)

Deep neural networks (DNNs) are typically optimized using various forms of mini-batch gradient descent algorithm. A major motivation for mini-batch gradient descent is that with a suitably chosen batch size, available computing resources can be optimally utilized (including parallelization) for fast model training. However, many works report the progressive loss of model generalization when the training batch size is increased beyond some limits. This is a scenario commonly referred to as generalization gap. Although several works have proposed different methods for alleviating the generalization gap problem, a unanimous account for understanding generalization gap is still lacking in the literature. This is especially important given that recent works have observed that several proposed solutions for generalization gap problem such learning rate scaling and increased training budget do not indeed resolve it. As such, our main exposition in this paper is to investigate and provide new perspectives for the source of generalization loss for DNNs trained with a large batch size. Our analysis suggests that large training batch size results in increased near-rank loss of units' activation (i.e. output) tensors, which consequently impacts model optimization and generalization. Extensive experiments are performed for validation on popular DNN models such as VGG-16, residual network (ResNet-56) and LeNet-5 using CIFAR-10, CIFAR-100, Fashion-MNIST and MNIST datasets.