AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

On Multi-Layer Basis Pursuit, Efficient Algorithms and Convolutional Neural Networks

Jeremias Sulam, Aviad Aberdam, Amir Beck, Michael Elad

TPAMI 2019 · Architectures and efficiency › RNNs and normalization · approach: Sparse dictionary representations · 111 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
unfold ML-ISTA iterations on top of the feed-forward CNN (same parameters) (ablation)ML-ISTA · CIFAR10accuracy (%): 79.00 -> 82.93Improved
unfold ML-FISTA iterations instead of feed-forwardCNN · SVHNaccuracy (%): 92.44 -> 93.79Improved
unfold ML-ISTA instead of feed-forwardCNN · MNISTaccuracy (%): 98.78 -> 99.10Improved
use global multi-layer pursuit (ML-ISTA) instead of layer-wise Layered BPunrolled CNN · CIFAR10accuracy (%): 80.73 -> 82.93Improved
use global multi-layer pursuit (ML-ISTA) instead of layer-wise Layered BPunrolled CNN · MNISTaccuracy (%): 99.19 -> 99.10Hurt
replace dictionaries with LISTA-style factors W and B (ML-LISTA)ML-ISTA · CIFAR10accuracy (%): 82.93 -> 82.68No clear effect
free all filters across unfoldings (All-Free) instead of tied ML-ISTAML-ISTA · CIFAR10accuracy (%): 82.93 -> 81.48Hurt
free all filters across unfoldings (All-Free) instead of tied ML-ISTAML-ISTA · SVHNaccuracy (%): 93.52 -> 94.06No clear effect
add intermediate-layer penalty (lambda_1 > 0) to basis pursuit (ablation)ISTA/ML-ISTA · synthetic two-layer model (SNR=10)recovery error: lower recovery error in both gamma_2 and gamma_1Improved
decrease mu in ML-ISTA/ML-FISTA (ablation)ML-ISTA/ML-FISTA · synthetic two-layer modelsolution accuracyMixed

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