math.CONov 16, 2026

Canonical activation patterns and ReLU pattern varieties

Authors: Bella Finkel, Yaoying Fu, Chenxi Wu

Abstract

A ReLU pattern variety records the polynomial relations among the coefficients of the affine maps associated with a prescribed set of activation patterns of a ReLU network. We study how the combinatorics of the canonical set of activation patterns of an unbiased ReLU network is reflected in the geometry of its pattern variety. In particular, we introduce the notion of pattern stability and characterize pattern stability and combinatorial stability for shallow networks. When the associated wall graph is connected and contains an edge of every neuron label, we resolve a problem of Alexandr and Montúfar by determining the defining ideal of the associated pattern variety explicitly. We establish a hierarchy of progressively weaker equivalence relations relating labeled combinatorial equivalence, equality and permutation equivalence of canonical pattern sets, and equivalences of the associated varieties. Finally, we relate functional dimension to the differential of the pattern map and show that it is bounded above by the dimension of the canonical pattern variety at every depth. For nondegenerate shallow networks, we give an exact formula for the functional dimension; for every shallow architecture, it equals the dimension of the canonical pattern variety on an open dense set of full Lebesgue measure.

Figures & tables

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

May 5, 2026

Most ReLU Networks Admit Identifiable Parameters

We study the realization map of deep ReLU networks, focusing on when a function determines its parameters up to scaling and permutation. To analyze hidden redundancies beyond these standard symmetries, we introduce a framework based on weighted polyhedral complexes. Our main result shows that for every architecture whose input and hidden layers have width at least two, there exists an open set of identifiable parameters. This implies that the functional dimension of every such architecture is exactly the number of parameters minus the number of hidden neurons. We further show that minimal functional representations can still have non-trivial parameter redundancies. Finally, we establish a generic depth hierarchy, whereby for an open set of parameters the realized function cannot be represented generically by any shallower network.
Sep 7, 2025

Toric geometry of ReLU neural networks

Given a continuous piecewise linear function f:Rn0→Rf:\mathbb{R}^{n_0} \to \mathbb{R} and a fixed architecture of feedforward ReLU neural networks, the exact function realization problem asks when a network of that architecture realizes ff. In contrast, expressivity concerns characterizing all functions exactly realizable by a given architecture. To address these questions systematically, we establish a connection between toric geometry and ReLU neural networks, allowing tools from algebraic geometry to analyze the structures. Starting with an unbiased ReLU network with rational weights, we define the ReLU fan, the ReLU toric variety, and the ReLU Cartier divisor associated with the network. This work also relates the tropical and the toric geometry: the former captures combinatorial data of ReLU networks, while the latter recovers the underlying geometric structure. As an application, we obtain a complete characterization of the functions realizable by unbiased shallow ReLU networks with rational weights. The characterization is expressed in terms of intersection numbers of ReLU Cartier divisors with torus-invariant curves and leads to a decomposition of piecewise-linear functions into shallow and non-shallow components.
Jan 30, 2026

Complete Identification of Deep ReLU Networks through Łukasiewicz Logic

Two deep ReLU networks can have entirely different architectures and parameters, yet realize the same function. We provide a complete characterization of this nonuniqueness. This is effected by building a symbolic calculus for deep ReLU networks, equivalence and simplification of networks becoming derivation of formulae, in close parallel to Shannon's analysis of switching circuits through Boolean logic. Inspired by Shannon, who turned circuit synthesis into the manipulation of Boolean formulae by the axioms of Boolean algebra, we turn ReLU network identification into the derivation of Łukasiewicz formulae by the axioms of many-valued (MV) logic. Two non-degenerate ReLU networks realize the same function on the unit cube if and only if one is obtained from the other by finitely many applications of the MV axioms for integer weights and biases, the divisible MV axioms for rational ones, and the Riesz MV axioms for real ones. The MV logic axioms characterize all symmetries of ReLU networks, the single-layer ones, which for tanh networks are the only kind, and the deep ones, spanning three or more layers. Our framework consists of three steps, an extraction algorithm turning a network into a substitution graph, whose represented formula has the network's input-output map as its truth function, a completeness theorem, by which functionally equivalent formulae are interderivable, and a construction algorithm returning from graphs to networks. The substitution graph is layered, carrying at each node a formula in the variables of the layer feeding it, encodes the network uniquely, and induces a new normal form for MV logic, compositional rather than flat as in the literature, hence retaining the algebraic structure of the network, with three local operations--node rewrite, layer collapse, layer expansion--realizing every derivation.