paper-with-me

Papers

Averaging Weights Leads to Wider Optima and Better Generalization

2018-03-14 · Pavel Izmailov, Dmitrii Podoprikhin, Timur Garipov, Dmitry Vetrov, Andrew Gordon Wilson

Deep neural networks are typically trained by optimizing a loss function with an SGD variant, in conjunction with a decaying learning rate, until convergence. We show that simple averaging of multiple points along the trajectory of SGD, with a cyclical or constant learning rate, leads to better generalization than conventional training. We also show that this Stochastic Weight Averaging (SWA) procedure finds much flatter solutions than SGD, and approximates the recent Fast Geometric Ensembling (FGE) approach with a single model. Using SWA we achieve notable improvement in test accuracy over conventional SGD training on a range of state-of-the-art residual networks, PyramidNets, DenseNets, and Shake-Shake networks on CIFAR-10, CIFAR-100, and ImageNet. In short, SWA is extremely easy to implement, improves generalization, and has almost no computational overhead.

📄 PDF Abstract BibTeX arXiv:1803.05407

Code (17)

timgaripov/swa 공식 구현 pytorch
ENSTA-U2IS-AI/torch-uncertainty pytorch
GeorgOhneH/WerbeSkip
Morxrc/kaggle_generate-dogs_public-1st-private-16th-solution
NajibYavari/DD2412 tf
benathi/fastswa-semi-sup pytorch
cmu-enyac/Renofeation pytorch
dice-group/aswa pytorch
greyhound101/IEEE-CIS-Fraud
izmailovpavel/contrib_swa_examples pytorch
izmailovpavel/torch_swa_examples pytorch
julianfaraone/SWA pytorch
kristpapadopoulos/keras-stochastic-weight-averaging
pycroscopy/atomai pytorch
simon-larsson/keras-swa tf
wjmaddox/swa_gaussian pytorch
zlwangustc/SWA_paddle paddle

Tasks

Image ClassificationStochastic Optimization

Methods 이 논문이 사용한 방법론

Concatenated Skip Connection A Concatenated Skip Connection is a type of skip connection that seeks to reuse features by concatenating them to new layers, allowing more information to be retained from…
Dense Block A Dense Block is a module used in convolutional neural networks that connects *all layers* (with matching feature-map sizes) directly with each other. It was originally…
XRP Customer Service Number +1-833-534-1729 설명 없음
Average Pooling 설명 없음
Zero-padded Shortcut Connection 설명 없음
Pyramidal Residual Unit A Pyramidal Residual Unit is a type of residual unit where the number of channels gradually increases as a function of the depth at which the layer occurs, which is similar to…
Pyramidal Bottleneck Residual Unit A Pyramidal Bottleneck Residual Unit is a type of residual unit where the number of channels gradually increases as a function of the depth at which the layer occurs, which is…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Similar Papers 제목 키워드 기반

Stochastic Weight Averaging Revisited

2022-01-03 · Hao Guo, Jiyong Jin, Bin Liu

Averaging neural network weights sampled by a backbone stochastic gradient descent (SGD) is a simple yet effective approach to assist the backbone SGD in finding better optima, in terms of generalization. From a statisti…

Two-Tailed Averaging: Anytime, Adaptive, Once-in-a-While Optimal Weight Averaging for Better Generalization

2022-09-26 · Gábor Melis

Tail Averaging improves on Polyak averaging's non-asymptotic behaviour by excluding a number of leading iterates of stochastic optimization from its calculations. In practice, with a finite number of optimization steps a…

Stochastic Optimization

Temporal Difference Weighted Ensemble For Reinforcement Learning

2019-09-25 · Takuma Seno, Michita Imai

Combining multiple function approximators in machine learning models typically leads to better performance and robustness compared with a single function. In reinforcement learning, ensemble algorithms such as an averagi…

reinforcement-learningReinforcement LearningReinforcement Learning (RL)

Optimal Mixture-of-Experts Model Averaging for Conditional Generative Models

2026-07-05 · Shijin Gong, Baihua He, Xinyu Zhang arxiv

Conditional generative models have emerged as powerful tools for sampling from target conditional distributions, driving substantial advances across a wide range of scientific and applied domains. As these models prolife…

Beyond Bayesian Model Averaging over Paths in Probabilistic Programs with Stochastic Support

2023-10-23 · Tim Reichelt, Luke Ong, Tom Rainforth

The posterior in probabilistic programs with stochastic support decomposes as a weighted sum of the local posterior distributions associated with each possible program path. We show that making predictions with this full…