Separation Capacity of Scattering Networks on Low-Dimensional Datasets

arXiv:2607.06048 · stat.ML, cs.IT, cs.LG, math.CA · Submitted 2026-07-07 · Read on arXiv

Konstantin Häberle, Helmut Bölcskei

stat.ML, cs.IT, cs.LG, math.CA

Submitted: 2026-07-07

Comments: 19 pages

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Trace — 3× in this paper · explained in 2 episodes
• Feature Extraction — 2× in this paper · explained in 3 episodes
• Intrinsic Dimension — 2× in this paper · explained in 3 episodes
• Generalization — 1× in this paper · explained in 14 episodes
• Representation Learning — 1× in this paper · explained in 3 episodes

Sources

• Separation Capacity of Scattering Networks
• Function-Counting Theory for Low-Dimensional Data Structures

Related papers

• Behavior of prediction performance metrics with rare events
• Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
• A Posterior-Dynamics Framework for Imaging Inverse Problems with Pretrained Diffusion Priors
• Scalable Posterior Uncertainty for Flexible Density-Based Clustering
• A Transport Geometry of Benign Overfitting
• Efficient quadratic entropy with distance sketches