AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

LassoNet: A Neural Network with Feature Sparsity

Ismael Lemhadri, Feng Ruan, Louis Abraham, Robert Tibshirani

JMLR 2021 · Architectures and efficiency › RNNs and normalization · approach: Representation geometry and collapse · 107 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 LassoNet instead of Fisher scorefeature selection (k=50) · MNIST-Fashion, NN learnertest accuracy: 0.671 -> 0.800Improved
use LassoNet instead of Fisher scorefeature selection (k=50) · ISOLET, NN learnertest accuracy: 0.793 -> 0.885Improved
use LassoNet instead of PFAfeature selection (k=50) · Activity, NN learnertest accuracy: 0.779 -> 0.849Improved
use LassoNet instead of HSIC-Lassofeature selection (k=50) · COIL-20, NN learnertest accuracy: 0.972 -> 0.991Improved
use LassoNet instead of PFAfeature selection (k=50) · MNIST, NN learnertest accuracy: 0.873 -> 0.873No clear effect
use LassoNet instead of HSIC-Lassofeature selection (k=50) · Mice Protein, NN learnertest accuracy: 0.958 -> 0.958No clear effect
use LassoNet instead of PFAfeature selection (k=50) · MNIST, tree learnertest accuracy: 0.879 -> 0.892Improved
use LassoNet instead of HSIC-Lassofeature selection (k=50) · ISOLET, tree learnertest accuracy: 0.888 -> 0.891No clear effect
use LassoNet instead of Fisher scorefeature selection (k=50) · COIL-20, tree learnertest accuracy: 0.996 -> 0.993Hurt
use LassoNet instead of HSIC-Lassofeature selection (k=50) · Activity, tree learnertest accuracy: 0.845 -> 0.860Improved
use 50 selected features instead of all featuresLassoNet · MNIST, NN learnertest accuracy: 0.928 -> 0.873Hurt
use LassoNet instead of Soft-Impute for imputationmatrix completion · MICE Protein imputationtest MSE: about 30% reductionImproved

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