[Paper Review] ZX-Calculus: Cyclotomic Supplementarity and Incompleteness for Clifford+T quantum mechanics

Emmanuel Jeandel, Simon Perdrix+2 more|arXiv (Cornell University)|Feb 7, 2017

Quantum Computing Algorithms and Architecture3 citations

TL;DR

This paper introduces cyclotomic supplementarity—a generalization of the supplementarity rule in the ZX-Calculus that unifies n subdiagrams with angles equally dividing the circle. It proves that for odd prime n, this rule is not derivable from existing axioms, establishing incompleteness of the Clifford+T fragment and proposing a new, complete axiomatization including an infinite family of cyclotomic supplementarity rules and a revised scalar rule.

ABSTRACT

The ZX-Calculus is a powerful graphical language for quantum mechanics and quantum information processing. The completeness of the language -- i.e. the ability to derive any true equation -- is a crucial question. In the quest of a complete ZX-calculus, supplementarity has been recently proved to be necessary for quantum diagram reasoning (MFCS 2016). Roughly speaking, supplementarity consists in merging two subdiagrams when they are parameterized by antipodal angles. We introduce a generalised supplementarity -- called cyclotomic supplementarity -- which consists in merging n subdiagrams at once, when the n angles divide the circle into equal parts. We show that when n is an odd prime number, the cyclotomic supplementarity cannot be derived, leading to a countable family of new axioms for diagrammatic quantum reasoning.We exhibit another new simple axiom that cannot be derived from the existing rules of the ZX-Calculus, implying in particular the incompleteness of the language for the so-called Clifford+T quantum mechanics. We end up with a new axiomatisation of an extended ZX-Calculus, including an axiom schema for the cyclotomic supplementarity.

Motivation & Objective

• To address the long-standing open problem of completeness for the Clifford+T fragment of quantum mechanics in the ZX-Calculus.
• To generalize the supplementarity rule to n-partite merging of diagrams with angles dividing the circle equally, introducing cyclotomic supplementarity.
• To demonstrate that this generalized rule cannot be derived when n is an odd prime, implying incompleteness of existing ZX-Calculus rules for the Clifford+T fragment.
• To propose a new, complete axiomatization of the ZX-Calculus by adding cyclotomic supplementarity and a revised scalar rule, replacing obsolete rules.

Proposed method

• Introduces cyclotomic supplementarity as a rule that unifies n subdiagrams when their angles partition the unit circle into n equal arcs.
• Uses matrix-based interpretation and algebraic number theory to show that certain scalar equalities (e.g., involving arccos(√(2/3))) are not derivable from existing rules.
• Applies a modified version of the Schr"oder de Witt and Zamdzhiev proof technique, using angle-multiplying interpretations to construct an infinite family of diagrams.
• Employs Cohn's irreducibility criterion to prove that specific angles (e.g., α₀ = ±π/2 ± arccos(√(2/3))) are not rational multiples of π, implying non-derivability.
• Constructs a family of interpretations J·K♯_{k,l} that preserve all rules except cyclotomic supplementarity, used to show independence of the new rules.
• Replaces the inverse and zero rules with a new scalar rule and proves its equivalence to the former under the new axiom set.

Experimental results

Research questions

• RQ1 Can the supplementarity rule be generalized to unify n subdiagrams with angles equally spaced on the circle, and is this generalization derivable from existing ZX-Calculus rules?
• RQ2 Is the Clifford+T fragment of the ZX-Calculus complete, and if not, what new axioms are required to achieve completeness?
• RQ3 For which values of n is the cyclotomic supplementarity rule independent of the existing ZX-Calculus rules?
• RQ4 Can the new scalar rule replace the inverse and zero rules while preserving completeness?
• RQ5 Does the addition of cyclotomic supplementarity rules restore completeness to the general ZX-Calculus?

Key findings

• Cyclotomic supplementarity for odd prime n cannot be derived from existing ZX-Calculus rules, proving that the language is incomplete for the Clifford+T fragment.
• The paper identifies a specific scalar equality involving arccos(√(2/3)) that is matrix-true but not derivable diagrammatically, demonstrating incompleteness.
• The angle α₀ = ±π/2 ± arccos(√(2/3)) is not a rational multiple of π, as shown by proving the irreducibility of the polynomial 3X⁴ + 2X² + 3 over ℤ.
• An infinite family of new axioms—cyclotomic supplementarity rules for all odd primes n—is required to achieve completeness for the Clifford+T fragment.
• The new axiomatization, including the revised scalar rule and cyclotomic supplementarity, restores completeness to the general ZX-Calculus.
• The proof of incompleteness is adapted using angle-multiplying interpretations J·K♯_{k,l} and an infinite set S of integers, showing that D1 has finitely many D2-form decompositions but infinitely many distinct interpretations under S.