AI Paper Atlas

Research map / Optimization / Optimizers

On permutation symmetries in Bayesian neural network posteriors: a variational perspective

Simone Rossi, Ankit Singh, Thomas Hannagan

NeurIPS 2023 · Optimization › Optimizers · approach: Differentiating through embedded optimization · 0 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
align the second VI posterior to the first via permutation matching before linear interpolation (ablation)Gaussian VI BNN (MLP, 3 layers) · MNIST / CIFAR10marginalized predictive-likelihood barrierImproved
align the second VI posterior to the first via permutation matching before linear interpolation (ablation)Gaussian VI BNN (ResNet20) · CIFAR10marginalized predictive-likelihood barrierImproved
increase network width with alignment (MLP and ResNet20)VI with alignment · MLP / ResNet20 on CIFAR10likelihood barrierImproved
increase ResNet20 width without alignmentnaive VI interpolation · ResNet20 on CIFAR10likelihood barrierNo clear effect
increase prior variance with alignmentVI with alignment · MLP on MNIST/Fashion-MNIST, ResNet20 x8 on CIFAR10likelihood barrierNo clear effect
increase prior variance without alignment (more diffuse prior)naive VI interpolation · MNIST, Fashion-MNIST, CIFAR10likelihood barrierImproved
increase prior varianceVI BNN · MNIST, Fashion-MNIST, CIFAR10train-test barrier gapHurt
cold posterior (T<1) versus warm posterior (T>1) with alignmentVI with alignment · MLP/ResNet20likelihood barrierImproved
replace VI with SGD MAP plus weight matching, narrow modelsResNet20 · CIFAR10 narrow widthsloss barrierNo clear effect
replace VI with SGD MAP plus weight matching, wide modelsResNet20 · CIFAR10 wide widthsloss barrierImproved
replace FRN with LayerNorm (ablation)BNN ResNet20 with alignment · CIFAR10likelihood barrierImproved
train with data augmentation (ablation)VI with alignment · CIFAR10likelihood barrierNo clear effect

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