Pith. sign in

REVIEW 1 cited by

Pure and Spurious Critical Points: a Geometric Study of 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 1910.01671 v2 pith:ZGM5JPI2 submitted 2019-10-03 cs.LG math.AGstat.ML

classification cs.LGmath.AGstat.ML
keywords linearspacelosscriticalfunctionalnetworksarchitecturesdeterminantal
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The critical locus of the loss function of a neural network is determined by the geometry of the functional space and by the parameterization of this space by the network's weights. We introduce a natural distinction between pure critical points, which only depend on the functional space, and spurious critical points, which arise from the parameterization. We apply this perspective to revisit and extend the literature on the loss function of linear neural networks. For this type of network, the functional space is either the set of all linear maps from input to output space, or a determinantal variety, i.e., a set of linear maps with bounded rank. We use geometric properties of determinantal varieties to derive new results on the landscape of linear networks with different loss functions and different parameterizations. Our analysis clearly illustrates that the absence of "bad" local minima in the loss landscape of linear networks is due to two distinct phenomena that apply in different settings: it is true for arbitrary smooth convex losses in the case of architectures that can express all linear maps ("filling architectures") but it holds only for the quadratic loss when the functional space is a determinantal variety ("non-filling architectures"). Without any assumption on the architecture, smooth convex losses may lead to landscapes with many bad minima.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Impact of Bottleneck Layers and Skip Connections on the Generalization of Linear Denoising Autoencoders

    stat.ML 2025-05 conditional novelty 7.0 of 10

    Two-layer linear denoising autoencoders show a bias-variance trade-off in bottleneck width, and skip connections reduce variance near the interpolation peak.

Pith tools