Rigorous Asymptotics for First-Order Algorithms Through the Dynamical Cavity Method
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914395642134528 |
|---|---|
| author | Dandi, Yatin Gamarnik, David Pernice, Francisco Zdeborová, Lenka |
| author_facet | Dandi, Yatin Gamarnik, David Pernice, Francisco Zdeborová, Lenka |
| contents | Dynamical Mean Field Theory (DMFT) provides an asymptotic description of the dynamics of macroscopic observables in certain disordered systems. Originally pioneered in the context of spin glasses by Sompolinsky and Zippelius (1982), it has since been used to derive asymptotic dynamical equations for a wide range of models in physics, high-dimensional statistics and machine learning. One of the main tools used by physicists to obtain these equations is the dynamical cavity method, which has remained largely non-rigorous. In contrast, existing mathematical formalizations have relied on alternative approaches, including Gaussian conditioning, large deviations over paths, or Fourier analysis. In this work, we formalize the dynamical cavity method and use it to give a new proof of the DMFT equations for General First Order Methods, a broad class of dynamics encompassing algorithms such as Gradient Descent and Approximate Message Passing. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_14573 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Rigorous Asymptotics for First-Order Algorithms Through the Dynamical Cavity Method Dandi, Yatin Gamarnik, David Pernice, Francisco Zdeborová, Lenka Disordered Systems and Neural Networks Machine Learning Probability Dynamical Mean Field Theory (DMFT) provides an asymptotic description of the dynamics of macroscopic observables in certain disordered systems. Originally pioneered in the context of spin glasses by Sompolinsky and Zippelius (1982), it has since been used to derive asymptotic dynamical equations for a wide range of models in physics, high-dimensional statistics and machine learning. One of the main tools used by physicists to obtain these equations is the dynamical cavity method, which has remained largely non-rigorous. In contrast, existing mathematical formalizations have relied on alternative approaches, including Gaussian conditioning, large deviations over paths, or Fourier analysis. In this work, we formalize the dynamical cavity method and use it to give a new proof of the DMFT equations for General First Order Methods, a broad class of dynamics encompassing algorithms such as Gradient Descent and Approximate Message Passing. |
| title | Rigorous Asymptotics for First-Order Algorithms Through the Dynamical Cavity Method |
| topic | Disordered Systems and Neural Networks Machine Learning Probability |
| url | https://arxiv.org/abs/2603.14573 |