AI Paper Atlas

Research map / Architectures and efficiency / CNN design

A Discriminative Gaussian Mixture Model with Sparsity

Hideaki Hayashi, Seiichi Uchida

ICLR 2021 · Architectures and efficiency › CNN design · approach: Theory of convolutional inductive bias · 0 citations (OpenAlex)

Read the paper

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 SDGM instead of the compared kernel Bayesian classifiersSDGM · Ripley, Waveform, Banana, Titanic, Breast Cancererror rate (%): 9.1 vs 9.1; 10.1 vs 9.6; 10.6 vs 10.5; 22.6 vs 22.6; 29.4 vs 29.9Mixed
Use sparse Bayesian learning to obtain sparse weightsSDGM · Ripley, Waveform, Banana, Titanic, Breast Cancernumber of nonzero weights: SDGM vs RVM: 6 vs 4; 11.0 vs 14.6; 11.1 vs 11.4; 74.5 vs 65.3; 15.7 vs 6.3Mixed
Replace CNN softmax with SDGMCNN · MNIST (D=2)error rate (%): 3.19 -> 1.81Improved
Replace CNN softmax with SDGMCNN · MNIST (D=10)error rate (%): 1.01 -> 0.86Improved
Replace CNN softmax with SDGMCNN · Fashion-MNISTerror rate (%): 8.78 -> 8.05Improved
Replace CNN softmax with SDGMDenseNet-40 · CIFAR-10error rate (%): 11.07 -> 10.46Improved
Replace CNN softmax with SDGMDenseNet-40 · CIFAR-100error rate (%): 22.99 -> 21.32Improved
Replace MobileNet softmax with SDGMMobileNet · ImageNeterror rate (%): 33.45 -> 32.40Improved
Add sparse Bayesian learning to the discriminative GMM (ablation)CNN classifier · MNIST (D=10), CIFAR-10error rate (%): 0.72 -> 0.86; 10.05 -> 10.46Hurt
Add sparse Bayesian learning to the discriminative GMM (ablation)CNN classifier · MNIST (D=2), Fashion-MNIST, CIFAR-100, ImageNeterror rate (%): 2.43 -> 1.81; 8.30 -> 8.05; 21.93 -> 21.32; 34.75 -> 32.40Improved
Increase the initial number of mixture components (ablation)SDGM · Synthetic dataerror rateImproved
Increase initial components beyond eight (ablation)SDGM · Synthetic datatest error rateNo 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