A note on regularised NTK dynamics with ...
Document type :
Pré-publication ou Document de travail
Title :
A note on regularised NTK dynamics with an application to PAC-Bayesian training
Author(s) :
Clerico, Eugenio [Auteur]
Guedj, Benjamin [Auteur]
MOdel for Data Analysis and Learning [MODAL]
The Inria London Programme [Inria-London]
The Alan Turing Institute
Inria Lille - Nord Europe
Institut National de Recherche en Informatique et en Automatique [Inria]
Department of Computer science [University College of London] [UCL-CS]
University College of London [London] [UCL]
Guedj, Benjamin [Auteur]

MOdel for Data Analysis and Learning [MODAL]
The Inria London Programme [Inria-London]
The Alan Turing Institute
Inria Lille - Nord Europe
Institut National de Recherche en Informatique et en Automatique [Inria]
Department of Computer science [University College of London] [UCL-CS]
University College of London [London] [UCL]
Publication date :
2023-12-20
HAL domain(s) :
Statistiques [stat]/Machine Learning [stat.ML]
Informatique [cs]/Apprentissage [cs.LG]
Informatique [cs]/Apprentissage [cs.LG]
English abstract : [en]
We establish explicit dynamics for neural networks whose training objective has a regularising term that constrains the parameters to remain close to their initial value. This keeps the network in a lazy training regime, ...
Show more >We establish explicit dynamics for neural networks whose training objective has a regularising term that constrains the parameters to remain close to their initial value. This keeps the network in a lazy training regime, where the dynamics can be linearised around the initialisation. The standard neural tangent kernel (NTK) governs the evolution during the training in the infinite-width limit, although the regularisation yields an additional term appears in the differential equation describing the dynamics. This setting provides an appropriate framework to study the evolution of wide networks trained to optimise generalisation objectives such as PAC-Bayes bounds, and hence potentially contribute to a deeper theoretical understanding of such networks.Show less >
Show more >We establish explicit dynamics for neural networks whose training objective has a regularising term that constrains the parameters to remain close to their initial value. This keeps the network in a lazy training regime, where the dynamics can be linearised around the initialisation. The standard neural tangent kernel (NTK) governs the evolution during the training in the infinite-width limit, although the regularisation yields an additional term appears in the differential equation describing the dynamics. This setting provides an appropriate framework to study the evolution of wide networks trained to optimise generalisation objectives such as PAC-Bayes bounds, and hence potentially contribute to a deeper theoretical understanding of such networks.Show less >
Language :
Anglais
Collections :
Source :
Files
- document
- Open access
- Access the document
- 2312.13259.pdf
- Open access
- Access the document
- 2312.13259
- Open access
- Access the document