Gradient descent on infinitely wide neural networks: global convergence and generalization (Q6200217)

From MaRDI portal
scientific article; zbMATH DE number 7822598
Language Label Description Also known as
English
Gradient descent on infinitely wide neural networks: global convergence and generalization
scientific article; zbMATH DE number 7822598

    Statements

    Gradient descent on infinitely wide neural networks: global convergence and generalization (English)
    0 references
    0 references
    0 references
    22 March 2024
    0 references
    Summary: Many supervised machine learning methods are naturally cast as optimization problems. For prediction models which are linear in their parameters, this often leads to convex problems for which many mathematical guarantees exist. Models which are nonlinear in their parameters such as neural networks lead to nonconvex optimization problems for which guarantees are harder to obtain. In this paper, we consider two-layer neural networks with homogeneous activation functions where the number of hidden neurons tends to infinity, and show how qualitative convergence guarantees may be derived. For the entire collection see [Zbl 07816361].
    0 references
    machine learning
    0 references
    neural networks
    0 references
    gradient descent
    0 references
    gradient flow
    0 references
    optimal transport
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references