English
Related papers

Related papers: Subsampling is not Magic: Why Large Batch Sizes Wo…

200 papers

Recent advances in the theoretical understanding of SGD led to a formula for the optimal batch size minimizing the number of effective data passes, i.e., the number of iterations times the batch size. However, this formula is of no…

Machine Learning · Computer Science 2021-11-22 Motasem Alfarra , Slavomir Hanzely , Alyazeed Albasyoni , Bernard Ghanem , Peter Richtarik

Modern deep neural network training is typically based on mini-batch stochastic gradient optimization. While the use of large mini-batches increases the available computational parallelism, small batch training has been shown to provide…

Machine Learning · Computer Science 2018-04-23 Dominic Masters , Carlo Luschi

Recent works have shown that high probability metrics with stochastic gradient descent (SGD) exhibit informativeness and in some cases advantage over the commonly adopted mean-square error-based ones. In this work we provide a formal…

Machine Learning · Computer Science 2022-11-03 Dragana Bajovic , Dusan Jakovetic , Soummya Kar

Modern deep neural networks often require distributed training with many workers due to their large size. As the number of workers increases, communication overheads become the main bottleneck in data-parallel minibatch stochastic gradient…

Machine Learning · Statistics 2024-11-07 Tim Tsz-Kit Lau , Weijian Li , Chenwei Xu , Han Liu , Mladen Kolar

We showcase important features of the dynamics of the Stochastic Gradient Descent (SGD) in the training of neural networks. We present empirical observations that commonly used large step sizes (i) lead the iterates to jump from one side of…

Machine Learning · Computer Science 2023-06-08 Maksym Andriushchenko , Aditya Varre , Loucas Pillaud-Vivien , Nicolas Flammarion

The choice of batch-size in a stochastic optimization algorithm plays a substantial role for both optimization and generalization. Increasing the batch-size used typically improves optimization but degrades generalization. To address the…

Machine Learning · Computer Science 2020-03-03 Yeming Wen , Kevin Luk , Maxime Gazeau , Guodong Zhang , Harris Chan , Jimmy Ba

The choice of batch sizes in minibatch stochastic gradient optimizers is critical in large-scale model training for both optimization and generalization performance. Although large-batch training is arguably the dominant training paradigm…

Machine Learning · Computer Science 2024-05-29 Tim Tsz-Kit Lau , Han Liu , Mladen Kolar

Deep learning networks are typically trained by Stochastic Gradient Descent (SGD) methods that iteratively improve the model parameters by estimating a gradient on a very small fraction of the training data. A major roadblock faced when…

Machine Learning · Computer Science 2020-06-11 Tao Lin , Lingjing Kong , Sebastian U. Stich , Martin Jaggi

For infinitesimal learning rates, stochastic gradient descent (SGD) follows the path of gradient flow on the full batch loss function. However moderately large learning rates can achieve higher test accuracies, and this generalization…

Machine Learning · Computer Science 2021-01-29 Samuel L. Smith , Benoit Dherin , David G. T. Barrett , Soham De

Applying Differentially Private Stochastic Gradient Descent (DPSGD) to training modern, large-scale neural networks such as transformer-based models is a challenging task, as the magnitude of noise added to the gradients at each iteration…

Machine Learning · Computer Science 2022-07-07 Ryuichi Ito , Seng Pei Liew , Tsubasa Takahashi , Yuya Sasaki , Makoto Onizuka

We propose a general yet simple theorem describing the convergence of SGD under the arbitrary sampling paradigm. Our theorem describes the convergence of an infinite array of variants of SGD, each of which is associated with a specific…

Machine Learning · Computer Science 2021-02-22 Robert Mansel Gower , Nicolas Loizou , Xun Qian , Alibek Sailanbayev , Egor Shulgin , Peter Richtarik

In value-based deep reinforcement learning with replay memories, the batch size parameter specifies how many transitions to sample for each gradient update. Although critical to the learning process, this value is typically not adjusted…

Machine Learning · Computer Science 2023-10-09 Johan Obando-Ceron , Marc G. Bellemare , Pablo Samuel Castro

The minibatch stochastic gradient descent method (SGD) is widely applied in deep learning due to its efficiency and scalability that enable training deep networks with a large volume of data. Particularly in the distributed setting, SGD is…

Machine Learning · Computer Science 2020-11-19 Qiwei Yuan , Weizhe Hua , Yi Zhou , Cunxi Yu

Mini-batch stochastic gradient descent and variants thereof have become standard for large-scale empirical risk minimization like the training of neural networks. These methods are usually used with a constant batch size chosen by simple…

Machine Learning · Computer Science 2017-06-29 Lukas Balles , Javier Romero , Philipp Hennig

This paper theoretically reanalyzes the convergence of the mini-batch stochastic gradient descent (SGD) for a structured minimization problem involving a finite-sum function with its gradient being stochastically approximated, and an…

Optimization and Control · Mathematics 2026-04-07 Runze Li , Jintao Xu , Wenxun Xing

Large-batch stochastic gradient descent (SGD) is widely used for training in distributed deep learning because of its training-time efficiency, however, extremely large-batch SGD leads to poor generalization and easily converges to sharp…

Machine Learning · Computer Science 2019-06-27 Kosuke Haruki , Taiji Suzuki , Yohei Hamakawa , Takeshi Toda , Ryuji Sakai , Masahiro Ozawa , Mitsuhiro Kimura

Differentially private SGD (DP-SGD) is one of the most popular methods for solving differentially private empirical risk minimization (ERM). Due to its noisy perturbation on each gradient update, the error rate of DP-SGD scales with the…

Machine Learning · Computer Science 2021-04-27 Yingxue Zhou , Zhiwei Steven Wu , Arindam Banerjee

Stochastic gradient descent updates parameters with summation gradient computed from a random data batch. This summation will lead to unbalanced training process if the data we obtained is unbalanced. To address this issue, this paper takes…

Machine Learning · Computer Science 2019-05-22 Tao Yi , Xingxuan Wang

Classical differential private DP-SGD implements individual clipping with random subsampling, which forces a mini-batch SGD approach. We provide a general differential private algorithmic framework that goes beyond DP-SGD and allows any…

Machine Learning · Computer Science 2023-07-26 Marten van Dijk , Phuong Ha Nguyen , Toan N. Nguyen , Lam M. Nguyen

We say an algorithm is batch size-invariant if changes to the batch size can largely be compensated for by changes to other hyperparameters. Stochastic gradient descent is well-known to have this property at small batch sizes, via the…

Machine Learning · Computer Science 2023-03-28 Jacob Hilton , Karl Cobbe , John Schulman