Complexity of Contextuality

TL;DR

This paper investigates the computational complexity of determining and finding noncontextual ontological models in generalized contextuality, revealing exponential bounds and fundamental differences in minimal model sizes.

Contribution

It introduces an algorithm for deciding the existence of noncontextual models of a given dimension and analyzes its complexity, highlighting the computational challenges involved.

Findings

Deciding noncontextual models is at least exponential in the theory's dimension.

Computing the smallest noncontextual model is generally NP-hard.

Explicit example shows minimal models can differ in size (5 vs 4).

Abstract

Generalized contextuality is a hallmark of nonclassical theories like quantum mechanics. Yet, three fundamental computational problems concerning its decidability and complexity remain open. First, determining the complexity of deciding if a theory admits a noncontextual ontological model; Second, determining the complexity of deciding if such a model is possible for a specific dimension $k$; Third, efficiently computing the smallest such model when it exists, given that finding the smallest ontological model is NP-hard. We address the second problem by presenting an algorithm derived from a geometric formulation and its reduction to the intermediate simplex problem in computational geometry. We find that the complexity of deciding the existence of a noncontextual ontological model of dimension $k$ is at least exponential in the dimension of the theory and at most exponential in $k$. This, in turn, implies that computing the smallest noncontextual ontological model is inefficient in general. Finally, we demonstrate the fundamental difference between finding the smallest noncontextual ontological model and the smallest ontological model using an explicit example wherein the respective minimum ontic sizes are five and four.