AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Hyper-Representations as Generative Models: Sampling Unseen Neural Network Weights

Konstantin Schürholt, Boris Knyazev, Xavier Giró-i-Nieto, Damian Borth

NeurIPS 2022 · Label-efficient and robust learning › Knowledge distillation · approach: Patient consistent distillation · 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
add layer-wise loss normalization to the autoencoder (ablation)hyper-representation + KDE30 sampling · SVHN, epoch 0test accuracy (%): 10.1 -> 51.5Improved
add layer-wise loss normalization (ablation)hyper-representation + KDE30 · CIFAR-10, epoch 0test accuracy (%): 15.5 -> 26.9Improved
add layer-wise loss normalization (ablation)hyper-representation + KDE30 · STL-10, epoch 0test accuracy (%): 12.7 -> 19.7Improved
add layer-wise loss normalization (ablation)hyper-representation + KDE30 · MNIST, epoch 0test accuracy (%): 63.2 -> 68.6Improved
add layer-wise loss normalization, after 25 epochs of fine-tuning (ablation)hyper-representation + KDE30 · SVHN, epoch 25test accuracy (%): 75.4 -> 74.2No clear effect
initialize with sampled weights and fine-tune 25 epochs instead of training from scratchS_KDE30 · MNISTtest accuracy (%): 83.3 -> 93.0Improved
initialize with sampled weights, 25 epochs vs scratch 50 epochsS_KDE30 · SVHNtest accuracy (%): 70.7 -> 74.2Improved
ensemble 15 sampled models instead of single scratch-trained model (50 epochs)S_KDE30 · SVHNtest accuracy (%): 70.7 -> 77.6Improved
use top-30% sampled population instead of 100% (ablation)KDE sampling · MNISTtest accuracy (%)Improved
use distribution-based sampling (KDE, GAN) instead of neighborhood sampling (ablation)sampling · MNISTtest accuracy (%)Improved
transfer SVHN to MNIST with KDE30 samples vs scratch training, epoch 1S_KDE30 · SVHN to MNISTtest accuracy (%): 20.6 -> 86.9Improved
S_KDE30 vs pre-train and fine-tune B_F, epoch 1S_KDE30 · SVHN to MNISTtest accuracy (%): 84.4 -> 86.9Improved

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