Abstract:Researchers are actively trying to gain better insights into the representational properties of convolutional neural networks for guiding better network designs and for interpreting a network's computational nature. Gaining such insights can be an arduous task due to the number of parameters in a network and the complexity of a network's architecture. Current approaches of neural network interpretation include Bayesian probabilistic interpretations and information theoretic interpretations. In this study, we take a different approach to studying convolutional neural networks by proposing an abstract algebraic interpretation using finite transformation semigroup theory. Specifically, convolutional layers are broken up and mapped to a finite space. The state space of the proposed finite transformation semigroup is then defined as a single element within the convolutional layer, with the acting elements defined by surrounding state elements combined with convolution kernel elements. Generators of the finite transformation semigroup are defined to complete the interpretation. We leverage this approach to analyze the basic properties of the resulting finite transformation semigroup to gain insights on the representational properties of convolutional neural networks, including insights into quantized network representation. Such a finite transformation semigroup interpretation can also enable better understanding outside of the confines of fixed lattice data structures, thus useful for handling data that lie on irregular lattices. Furthermore, the proposed abstract algebraic interpretation is shown to be viable for interpreting convolutional operations within a variety of convolutional neural network architectures.

Lie Group Algebra Convolutional Filters

Lie Group Algebra Convolutional Filters

Lie Algebra Convolutional Neural Networks with Automatic Symmetry Extraction

Convolutional Filtering with RKHS Algebras

Simplicial Convolutional Filters

Warped Convolutional Networks: Bridge Homography to sl(3) algebra by Group Convolution

Higher-Order Nonlinear Complementary Filtering on Lie Groups

Non Commutative Convolutional Signal Models in Neural Networks: Stability to Small Deformations

Lie Algebra Canonicalization: Equivariant Neural Operators under arbitrary Lie Groups

A Construction of the Lie Algebra of a Lie Group in Isabelle/HOL

Lie Algebraic Method in Nonlinear Filtering with State Space Dimension Up to 6

B-Spline CNNs on Lie Groups

Seeing Convolution Through the Eyes of Finite Transformation Semigroup Theory: An Abstract Algebraic Interpretation of Convolutional Neural Networks

Classification of All Finite-Dimensional Nonlinear Filters from Lie Algebraic Point of View: State Dimension 2.

Deep Learning Symmetries and Their Lie Groups, Algebras, and Subalgebras from First Principles

Enabling equivariance for arbitrary Lie groups

Discrete‐time invariant extended Kalman filter on matrix Lie groups

Lie Group Decompositions for Equivariant Neural Networks

Equivariant Filter Design for Kinematic Systems on Lie Groups

Generalized low-pass filters and multiresolution analyses

Equivariant Filters are Equivariant