Variational Bounds for Perceptron Learning from Structured Data

Francesco Camilli, Pierluigi Contucci, Federica Gerace, Emanuele Mingione

arXivarxiv/2608.04882Aug 5, 2026cs.LGscore 0

We introduce a variational approach to a finite-temperature continuous-spin perceptron trained on a Gaussian mixture. The model allows for a broad class of concave utilities and log-concave separable prior measures on the spins. By combining the interpolation method with log-concavity and concentration estimates, we derive lower and upper minimax variational bounds for the limiting quenched pressure. Remarkably, the two bounds differ only in the order of optimization of two variational parameters, while all remaining extrema are controlled by the concave--convex structure of the variational…

SourceJSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2608.04882
curl -X POST /api/v1/records/arxiv/2608.04882/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2608.04882/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'