Deep learning Gauss-Manin connections

DOI10.1007/S00006-022-01207-1zbMATH Open1487.14128arXiv2007.13786OpenAlexW4213284108WikidataQ114232361 ScholiaQ114232361MaRDI QIDQ2113261FDOQ2113261

Avinash Kulkarni, Kathryn Heal, Emre Can Sertöz

Publication date: 11 March 2022

Published in: Advances in Applied Clifford Algebras (Search for Journal in Brave)

Abstract: The Gauss-Manin connection of a family of hypersurfaces governs the change of the period matrix along the family. This connection can be complicated even when the equations defining the family look simple. When this is the case, it is computationally expensive to compute the period matrices of varieties in the family via homotopy continuation. We train neural networks that can quickly and reliably guess the complexity of the Gauss-Manin connection of a pencil of hypersurfaces. As an application, we compute the periods of 96% of smooth quartic surfaces in projective 3-space whose defining equation is a sum of five monomials; from the periods of these quartic surfaces, we extract their Picard numbers and the endomorphism fields of their transcendental lattices.

Full work available at URL: https://arxiv.org/abs/2007.13786

zbMATH Keywords

connectionsmachine-learningneural-networks

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Computational aspects of algebraic surfaces (14Q10) Variation of Hodge structures (algebro-geometric aspects) (14D07) Period matrices, variation of Hodge structure; degenerations (32G20) Picard groups (14C22)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• The Magma algebra system. I: The user language
• DGM: a deep learning algorithm for solving partial differential equations
• Factoring polynomials with rational coefficients
• Multilayer feedforward networks are universal approximators
• On the periods of certain rational integrals. I, II
• Approximation by superpositions of a sigmoidal function
• Hodge groups of K3 surfaces.
• On the differentiation of De Rham cohomology classes with respect to parameters
• Shortest paths algorithms: Theory and experimental evaluation
• Networks and the best approximation property
• Examples of \(K3\) surfaces with real multiplication
• Algebraic geometry over the complex numbers
• Computing periods of hypersurfaces
• Using machine learning to improve cylindrical algebraic decomposition
• On reconstructing subvarieties from their periods
• A Numerical Transcendental Method in Algebraic Geometry: Computation of Picard Groups and Related Invariants
• Zeta functions of nondegenerate hypersurfaces in toric varieties via controlled reduction in \(p\)-adic cohomology

Cited In (4)

• From the string landscape to the mathematical landscape: a machine-learning outlook
• Deep learning the hyperbolic volume of a knot
• Cluster algebras: network science and machine learning
• Is deep learning a useful tool for the pure mathematician?

Uses Software

• Magma
• SageMath
• DGM

Recommendations

• Topological Approaches to Deep Learning 👍 👎
• Mathematical Methods in Deep Learning 👍 👎
• Deep learning of conjugate mappings 👍 👎
• A singular Riemannian geometry approach to deep neural networks. I: Theoretical foundations 👍 👎
• Graph Laplacians, Riemannian manifolds, and their machine-learning 👍 👎
• Estimation of a regression function on a manifold by fully connected deep neural networks 👍 👎
• Deep Neural Networks in a Mathematical Framework 👍 👎
• Differential Geometry Methods for Constructing Manifold-Targeted Recurrent Neural Networks 👍 👎
• Mathematical Aspects of Deep Learning 👍 👎
• The Modern Mathematics of Deep Learning 👍 👎

This page was built for publication: Deep learning Gauss-Manin connections

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2113261)