Quantitative Monoidal Algebra: Axiomatising Distance with String Diagrams

TL;DR

This paper introduces a new quantitative framework for string diagram calculi, allowing for axiomatization of distances between diagrams in resource-sensitive domains like quantum and probabilistic systems.

Contribution

It develops a monoidal setting for quantitative theories, defining syntactic categories and models enriched over a quantale to measure diagram distances.

Findings

Framework applicable to probabilistic and linear systems

Enrichment over a quantale captures notions of distance

Extends diagrammatic reasoning to approximate equivalences

Abstract

String diagrammatic calculi have become increasingly popular in fields such as quantum theory, circuit theory, probabilistic programming, and machine learning, where they enable resource-sensitive and compositional algebraic analysis. Traditionally, the equations of diagrammatic calculi only axiomatise exact semantic equality. However, reasoning in these domains often involves approximations rather than strict equivalences. In this work, we develop a quantitative framework for diagrammatic calculi, where one may axiomatise notions of distance between string diagrams. Unlike similar approaches, such as the quantitative theories introduced by Mardare et al., this requires us to work in a monoidal rather than a cartesian setting. We define a suitable notion of monoidal theory, the syntactic category it freely generates, and its models, where the concept of distance is established via enrichment over a quantale. To illustrate the framework, we provide examples from probabilistic and linear systems analysis.