Pith. sign in

REVIEW 11 cited by

Gradient descent aligns the layers of deep linear networks

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1810.02032 v2 pith:I7Q3R6PY submitted 2018-10-04 cs.LG math.OCstat.ML

Gradient descent aligns the layers of deep linear networks

classification cs.LG math.OCstat.ML
keywords gradientdescentlinearweightalignmentappliedconvergesdecreasing
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

This paper establishes risk convergence and asymptotic weight matrix alignment --- a form of implicit regularization --- of gradient flow and gradient descent when applied to deep linear networks on linearly separable data. In more detail, for gradient flow applied to strictly decreasing loss functions (with similar results for gradient descent with particular decreasing step sizes): (i) the risk converges to 0; (ii) the normalized i-th weight matrix asymptotically equals its rank-1 approximation $u_iv_i^{\top}$; (iii) these rank-1 matrices are aligned across layers, meaning $|v_{i+1}^{\top}u_i|\to1$. In the case of the logistic loss (binary cross entropy), more can be said: the linear function induced by the network --- the product of its weight matrices --- converges to the same direction as the maximum margin solution. This last property was identified in prior work, but only under assumptions on gradient descent which here are implied by the alignment phenomenon.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 11 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A Theory of Saddle Escape in Deep Nonlinear Networks

    cs.LG 2026-05 unverdicted novelty 8.0

    Derives exact Frobenius norm imbalance identity for deep nonlinear networks, classifies activations into four classes, and obtains critical-depth escape time law τ★ = Θ(ε^{-(r-2)}) from reduction to scalar ODE on perm...

  2. Conservation Laws from Data Symmetry in Neural Networks

    cs.LG 2026-06 unverdicted novelty 7.0

    Data symmetries generically do not induce conserved quantities in NN training for analytic non-polynomial losses, but can for MSE with tensorizable networks.

  3. A Theory of Saddle Escape in Deep Nonlinear Networks

    cs.LG 2026-05 unverdicted novelty 7.0

    Derives exact norm-imbalance identity for deep nonlinear nets, classifying activations into four classes and yielding escape time law τ★ = Θ(ε^{-(r-2)}) governed by bottleneck depth r.

  4. A Theory of Saddle Escape in Deep Nonlinear Networks

    cs.LG 2026-05 conditional novelty 7.0

    An exact norm-imbalance identity classifies activations into four classes and reduces deep nonlinear training flow to a scalar ODE that predicts saddle escape time scaling as ε to the power of minus (r-2) for r bottle...

  5. Implicit Bias in Deep Linear Discriminant Analysis

    cs.LG 2026-03 unverdicted novelty 7.0

    Gradient flow on deep diagonal linear LDA networks with balanced initialization converts additive updates to multiplicative updates, automatically conserving the (2/L) quasi-norm.

  6. A Theory on Flow Matching with Neural Networks

    cs.LG 2026-06 unverdicted novelty 6.0

    Establishes convergence guarantees for overparameterized 2-layer ReLU networks in flow matching, generalization bounds for the velocity-field objective, and Wasserstein guarantees for generated samples, using multi-ta...

  7. On the global convergence of gradient descent for wide shallow models with bounded nonlinearities

    math.OC 2026-05 unverdicted novelty 6.0

    Gradient descent on wide shallow models with bounded nonlinearities converges globally in the mean-field limit as non-global critical points are unstable under the dynamics.

  8. The Effect of Mini-Batch Noise on the Implicit Bias of Adam

    cs.LG 2026-02 unverdicted novelty 6.0

    Mini-batch noise reverses how Adam's β2 controls anti-regularization, making default momentum values suitable for small batches but requiring β1 closer to β2 for large batches to favor flatter minima.

  9. Prediction horizon shapes representations in predictive learning

    cs.LG 2025-11 unverdicted novelty 6.0

    Longer prediction horizons in predictive learning interact with model biases to recover the latent geometry of the task.

  10. Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models

    cs.LG 2024-01 unverdicted novelty 6.0

    SPIN lets weak LLMs become strong by self-generating training data from previous model versions and training to prefer human-annotated responses over its own outputs, outperforming DPO even with extra GPT-4 data on be...

  11. Conservation Laws for Modern Neural Architectures

    cs.LG 2026-06 unverdicted novelty 5.0

    Unified framework characterizes conservation laws for gradient flow in feedforward networks with GELU/SiLU/SwiGLU, multihead attention with positional encodings, and MoE models under various gating.