Pith. sign in

REVIEW

Sampling the space of solutions of an artificial neural network

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 2503.08266 v2 pith:ODRLR7RK submitted 2025-03-11 cond-mat.dis-nn cond-mat.stat-mechmath.PR

Sampling the space of solutions of an artificial neural network

classification cond-mat.dis-nn cond-mat.stat-mechmath.PR
keywords algorithmartificialflatlandscapelow-energymanifoldnetworknetworks
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 weight space of an artificial neural network can be systematically explored using tools from statistical mechanics. We employ a combination of a hybrid Monte Carlo algorithm which performs long exploration steps, a ratchet-based algorithm to investigate connectivity paths, and coupled replica models simulations to study subdominant flat regions. Our analysis focuses on one hidden layer networks and spans a range of energy levels and constrained density regimes. Near the interpolation threshold, the low-energy manifold shows a spiky topology. In the overparameterized regime, however, the low-energy manifold becomes entirely flat, forming an extended complex structure that is easy to sample. These numerical results are supported by an analytical study of the training error landscape, and we show numerically that the qualitative features of the loss landscape are robust across different data structures. Our study aims to provide new methodological insights for developing scalable methods for large networks.

discussion (0)

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