[Paper Review] A categorical model for a quantum circuit description language

Francisco Franklin Sousa Rios, Peter Selinger|arXiv (Cornell University)|Jun 8, 2017

Formal Methods in Verification20 citations

TL;DR

This paper introduces Proto-Quipper-M, a type-safe quantum circuit description language based on a categorical model of parameters and state in symmetric monoidal categories. It formalizes a fragment of Quipper with denotational and operational semantics, enabling safe, general-purpose description of quantum circuits and other morphisms in any symmetric monoidal category, with proven type safety, soundness, and adequacy.

ABSTRACT

Quipper is a practical programming language for describing families of quantum circuits. In this paper, we formalize a small, but useful fragment of Quipper called Proto-Quipper-M. Unlike its parent Quipper, this language is type-safe and has a formal denotational and operational semantics. Proto-Quipper-M is also more general than Quipper, in that it can describe families of morphisms in any symmetric monoidal category, of which quantum circuits are but one example. We design Proto-Quipper-M from the ground up, by first giving a general categorical model of parameters and state. The distinction between parameters and state is also known from hardware description languages. A parameter is a value that is known at circuit generation time, whereas a state is a value that is known at circuit execution time. After finding some interesting categorical structures in the model, we then define the programming language to fit the model. We cement the connection between the language and the model by proving type safety, soundness, and adequacy properties.

Motivation & Objective

• To formalize a safe, general-purpose language for describing families of quantum circuits and morphisms in symmetric monoidal categories.
• To distinguish between parameters (known at circuit generation time) and state (known at execution time), mirroring hardware description practices.
• To develop a categorical model that captures the essential structures of parameterized and stateful quantum circuits.
• To ensure the language is type-safe, sound, and adequate with respect to its denotational semantics.
• To generalize Quipper by abstracting beyond quantum circuits to any symmetric monoidal category.

Proposed method

• Design a categorical model of parameters and state using symmetric monoidal categories, with parameters as external values and state as internal circuit data.
• Define a formal syntax and type system for Proto-Quipper-M based on the categorical model, ensuring type safety from the outset.
• Establish denotational semantics via a functorial interpretation mapping language terms to morphisms in a symmetric monoidal category.
• Define operational semantics through a small-step reduction system for evaluating programs.
• Prove key metatheoretic properties: type safety, soundness (preservation of types under reduction), and adequacy (full abstraction between syntax and semantics).
• Use categorical constructions such as tensor products, internal homs, and natural transformations to model parameterized and stateful computation.

Experimental results

Research questions

• RQ1 How can a quantum circuit description language be made formally type-safe while preserving generality?
• RQ2 What categorical structures are necessary to model the distinction between parameters and state in circuit generation and execution?
• RQ3 Can a formal semantics be constructed for a fragment of Quipper that supports both denotational and operational reasoning?
• RQ4 How can the language be generalized beyond quantum circuits to other symmetric monoidal categories?
• RQ5 What metatheoretic properties—such as type safety and adequacy—can be proven for such a language?

Key findings

• Proto-Quipper-M is formally type-safe, ensuring that well-typed programs do not go wrong during execution.
• The language’s denotational semantics is sound, meaning that reduction preserves typing and semantic equivalence.
• Adequacy is proven, showing that the syntactic and semantic models fully capture each other’s behavior.
• The categorical model successfully captures the distinction between parameters and state, enabling correct modeling of circuit generation and execution.
• The language is general enough to describe families of morphisms in any symmetric monoidal category, not just quantum circuits.
• The formalization provides a foundation for building verified, high-level quantum programming languages with strong theoretical guarantees.