AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Vector-output ReLU Neural Network Problems are Copositive Programs: Convex Analysis of Two Layer Networks and Polynomial-time Algorithms

Arda Sahiner, Tolga Ergen, John M. Pauly, Mert Pilanci

ICLR 2021 · Optimization › Implicit bias and generalization · approach: Shallow ReLU network convergence · 3 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 soft-thresholded SVD under the stated whitened-data and singular-vector conditionsTwo-layer ReLU network · Whitened CIFAR-10 and CIFAR-100training objectiveNo clear effect
Use soft-thresholded SVD instead of SGDTwo-layer ReLU network · Whitened CIFAR-10solve time: 0.018 secondsImproved
Use soft-thresholded SVD instead of SGDTwo-layer ReLU network · Whitened CIFAR-100solve time: 0.36 secondsImproved
Increase hidden units from 10 to 50SGD-trained two-layer ReLU network · Synthetic Gaussian inputs with generated labelstraining loss relative to global optimumImproved
Replace ReLU activation with linear activation (ablation)Two-layer network on whitened CIFAR-10 continuous labels · CIFAR-10 test settest lossHurt
Use ReLU rather than the zero network at β=10^{-2} (ablation)Two-layer network on whitened CIFAR-10 continuous labels · CIFAR-10 test settest loss: 125383 -> 63038Improved
Use the copositive relaxationVector-output ReLU network · Spiral classificationtraining lossNo clear effect
Use the copositive relaxation instead of SGDVector-output ReLU network · Whitened MNIST and CIFAR-10 image fittingtraining objectiveNo 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