Pith. sign in

REVIEW 3 cited by

Provable Benefit of Orthogonal Initialization in Optimizing 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 2001.05992 v1 pith:HBMIPTT4 submitted 2020-01-16 cs.LG cs.NEmath.OCstat.ML

Provable Benefit of Orthogonal Initialization in Optimizing Deep Linear Networks

classification cs.LG cs.NEmath.OCstat.ML
keywords deepinitializationnetworksconvergenceorthogonaldepthefficientempirical
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

The selection of initial parameter values for gradient-based optimization of deep neural networks is one of the most impactful hyperparameter choices in deep learning systems, affecting both convergence times and model performance. Yet despite significant empirical and theoretical analysis, relatively little has been proved about the concrete effects of different initialization schemes. In this work, we analyze the effect of initialization in deep linear networks, and provide for the first time a rigorous proof that drawing the initial weights from the orthogonal group speeds up convergence relative to the standard Gaussian initialization with iid weights. We show that for deep networks, the width needed for efficient convergence to a global minimum with orthogonal initializations is independent of the depth, whereas the width needed for efficient convergence with Gaussian initializations scales linearly in the depth. Our results demonstrate how the benefits of a good initialization can persist throughout learning, suggesting an explanation for the recent empirical successes found by initializing very deep non-linear networks according to the principle of dynamical isometry.

discussion (0)

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

Forward citations

Cited by 3 Pith papers

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

  1. When Evidence is Sparse: Weakly Supervised Early Failure Alerting in Dialogs and LLM-Agent Trajectories

    cs.CL 2026-06 unverdicted novelty 7.0

    Proposes attention-based sparse evidence learning and a preference-conditioned stopping policy that improves early failure alerting Pareto frontiers by 3-42% over prior methods on five dialog and agent benchmarks.

  2. Criticality and Saturation in Orthogonal Neural Networks

    cs.LG 2026-05 conditional novelty 7.0

    Derives layer-wise recursions for finite-width tensors under orthogonal initialization that reproduce the observed large-depth stability of nonlinear networks.

  3. PLATO: Pointer Learner for Agent and Task Openness

    cs.AI 2026-07 conditional novelty 6.0

    A pointer-network actor plus GNN critic jointly handles agent and task openness in MARL without fixed bounds, with proofs of well-definedness and strong wildfire results.