Neural Network Symmetrisation in Concrete Settings

Cornish (2024) recently gave a general theory of neural network symmetrisation in the abstract context of Markov categories. We give a high-level overview of these results, and their concrete implications for the symmetrisation of deterministic functions and of Markov kernels.

cs.LG Neural Networks and Applications Infrastructure Maintenance and Monitoring Mathematics Computer Science

View graph Source

Paper

Similar papers

Stochastic Neural Network Symmetrisation in Markov CategoriesRob Cornish2024 · arXiv (Cornell University) · 7 citations In Library

On the Ability of Deep Networks to Learn Symmetries from Data: A Neural Kernel TheoryAndrea Perin, Stéphane Deny2024 · arXiv (Cornell University) · 9 citations In Library

A Functional Perspective on Learning Symmetric Functions with Neural NetworksAaron Zweig, Joan Bruna2020 · arXiv (Cornell University) · 25 citations In Library

Data Symmetries and Learning in Fully Connected Neural NetworksFabio Anselmi, Luca Manzoni, Alberto d’Onofrio et al.2023 · IEEE Access · 6 citations In Library

Symplectic Neural Networks Based on Dynamical SystemsBenjamin K. Tapley2024 · arXiv (Cornell University) · 6 citations In Library

Machine Learning SymmetryShailesh Lal2022 · 1 citation In Library

Probabilistic Symmetries and Invariant Neural NetworksBenjamin Bloem-Reddy, Yee Whye Teh2020 · Journal of Machine Learning Research In Library

Non-Linear Interactions in Neural Network Operators: New Theorems on Symmetry-Preserving TransformationsRômulo Damasclin Chaves dos Santos, Jorge Henrique de Oliveira Sales2025 · arXiv (Cornell University) · 0 citations In Library

Invariants for neural automataJone Uria–Albizuri, Giovanni Sirio Carmantini, Peter beim Graben et al.2023 · arXiv (Cornell University) In Library

Scale Equivariant Graph MetanetworksIoannis Kalogeropoulos, Giorgos Bouritsas, Yannis Panagakis2024 · Neural Information Processing Systems · 19 citations In Library

Patents near this paper

MULTI-LAYER NEURAL NETWORKS USING SYMMETRIC TENSORSUS 11,205,114 · INTEL CORPORATION · 2021 HIERARCHICAL SCALABLE NEUROMORPHIC SYNAPTRONIC SYSTEM FOR SYNAPTIC AND STRUCTURAL PLASTICITYUS 8,996,430 · INTERNATIONAL BUSINESS MACHINES CORPORATION · 2015 HIERARCHICAL SCALABLE NEUROMORPHIC SYNAPTRONIC SYSTEM FOR SYNAPTIC AND STRUCTURAL PLASTICITYUS 10,140,571 · INTERNATIONAL BUSINESS MACHINES CORPORATION · 2018 DISTRIBUTED STATE VIA CASCADES OF TENSOR DECOMPOSITIONS AND NEURON ACTIVATION BINDING ON NEUROMORPHIC HARDWAREUS 11,049,000 · INTERNATIONAL BUSINESS MACHINES CORPORATION · 2021 NEURAL NETWORK WITH NON-LINEAR TRANSFORMATIONSUS 4,979,126 · AI WARE INCORPORATED, A OH CORP. · 1990

Source of record

arXiv