AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Gradient descent aligns the layers of deep linear networks

Ziwei Ji, Matus Telgarsky

ICLR 2019 · Optimization › Implicit bias and generalization · approach: Explicit and implicit regularization interplay · 103 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
Apply gradient flow to a deep linear network on linearly separable dataDeep linear network · Linearly separable datarisk: risk converges to 0Improved
Apply gradient flow to a deep linear network on linearly separable dataDeep linear network · Linearly separable datanormalized weight matrix rank-one alignment: W_k/||W_k||_F converges to u_kv_k^TImproved
Apply gradient flow to a deep linear network on linearly separable dataDeep linear network · Linearly separable dataadjacent-layer alignment: |v_{k+1}^T u_k| -> 1Improved
Use logistic loss with gradient flowDeep linear network · Linearly separable datapredictor alignment with maximum-margin solution: wnn/||wnn|| -> u-barImproved
Use adaptive decreasing step sizes for gradient descentDeep linear network · Linearly separable datarisk: risk converges to 0Improved
Use adaptive decreasing step sizes for gradient descentDeep linear network · Linearly separable datanormalized network predictor alignment: normalized network predictor aligns with v_1Improved
Use logistic loss with gradient descentDeep linear network · Linearly separable datapredictor alignment with maximum-margin solution: wnn/∏||W_k||_F -> u-barImproved

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