AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Stochastic Anderson Mixing for Nonconvex Stochastic Optimization

Fuchao Wei, Chenglong Bao, Yang Liu

NeurIPS 2021 · Optimization › Nonconvex and smooth optimization · approach: Stochastic variance reduction · 4 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use AdaSAM instead of the strongest listed baselineResNet18, ResNet20, ResNet32, ResNet44, ResNet56, and WideResNet16-4 · CIFAR-10test top-1 accuracy (%): 94.82±.15 -> 95.17±.10; 92.07±.04 -> 92.43±.19; 92.86±.15 -> 93.22±.32; 93.26±.24 -> 93.57±.14; 93.76±.16 -> 93.77±.12; Improved
Use AdaSAM instead of the strongest listed baselineResNet18, ResNeXt50, and DenseNet121 · CIFAR-100test top-1 accuracy (%): 77.63±.35 -> 78.13±.14; 78.93±.12 -> 79.31±.27; 79.37±.16 -> 80.09±.52Improved
Add adaptive regularization and damped projection rather than use RAMAnderson mixing on CIFAR-10 and CIFAR-100 · CIFAR-10 and CIFAR-100test top-1 accuracy (%): 95.10±.05 -> 95.17±.10; 92.21±.09 -> 92.43±.19; 93.05±.43 -> 93.22±.32; 93.42±.13 -> 93.57±.14; 93.76±.16 -> 93.77±.12; Improved
Add variance reduction using SVRG (ablation)AdaSAM · MNIST trainingtraining loss: 10^{-9}Improved
Precondition AdaSAM with Adagrad or RMSprop (ablation)AdaSAM · MNIST trainingtraining loss: 10^{-3}Mixed
Precondition AdaSAM with Adam (ablation)AdaSAM on Penn TreeBank LSTM · Penn TreeBankperplexityImproved
Alternate vanilla SGD with AdaSAM every 5 steps (ablation)AdaSAM · CIFAR-100, ResNeXt50test top-1 accuracy (%): 78.41±.54 -> 79.47±.44Improved
Alternate Adam with AdaSAM every 5 steps (ablation)Adam · CIFAR-10 and CIFAR-100test top-1 accuracy (%)Mixed
Use alternating optimizer steps to reduce computational overhead (ablation)AdaSAM-SGD and AdaSAM-Adam · CIFAR-10 and CIFAR-100test top-1 accuracy (%): 79.31±.27 -> 79.47±.44Mixed
Use vanilla AdaSAM without a suitable preconditioner (ablation)AdaSAM · Penn TreeBank LSTMperplexityHurt
Use preconditioned mixing (ablation)AdaSAM · MNIST trainingtraining loss: 10^{-3}Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy