⚛️ quantum physics

Kirkwood-Dirac classical states based on discrete Fourier transform: Representation with directed graph

This paper investigates the structural characteristics of Kirkwood-Dirac classical states under discrete Fourier transform bases by proving their convex hull property in $p^r$ -dimensional spaces and introducing a directed graph framework to characterize these states in arbitrary dimensions, thereby generalizing and unifying previous results.

Original authors: Lin-Yan Cai, Ying-Hui Yang, Zhu-Jun Zheng

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Lin-Yan Cai, Ying-Hui Yang, Zhu-Jun Zheng

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The Quantum "Lie Detector"

Imagine you are trying to describe a secret recipe (a quantum state). In the classical world, recipes are straightforward lists of ingredients with positive amounts (e.g., "2 cups of flour, 1 egg"). You can never have "-2 cups of flour."

However, in the quantum world, things get weird. To describe a quantum recipe, scientists use a special tool called the Kirkwood-Dirac (KD) distribution. Think of this as a "quantum ingredient list." Sometimes, this list contains negative numbers or even imaginary numbers (like $\sqrt{-1}$ ).

KD-Classical: If your ingredient list only has normal, positive numbers, the recipe is "Classical." It behaves like the normal world.
KD-Nonclassical: If your list has negative or imaginary numbers, the recipe is "Nonclassical." This is where the magic happens—this is where quantum computers get their superpowers (like speed and security).

The big question scientists have been asking is: "If we mix together all the 'pure' classical recipes, can we make every possible 'mixed' classical recipe?"

In math terms: Is the set of all classical states just a "cloud" (convex hull) made by connecting all the pure classical states?

The Problem: A Puzzle with Missing Pieces

For a long time, scientists knew the answer was "Yes" for very specific, simple puzzle sizes (like prime numbers). But for other sizes (like the number 6), they found counter-examples where the answer was "No."

This paper tackles a specific, very important type of puzzle where the rules are based on the Discrete Fourier Transform (DFT). Think of the DFT as a specific way of shuffling the ingredients in your recipe. The authors wanted to know: If we shuffle ingredients using this specific DFT rule, does the "Yes" answer hold true for all sizes?

The Solution: Two New Tools

The authors, Lin-Yan Cai, Ying-Hui Yang, and Zhu-Jun Zheng, came up with two clever ways to solve this.

1. The "Prime Number" Proof (The Simple Case)

First, they looked at puzzle sizes that are powers of a single prime number (like $2^3 = 8$ or $3^2 = 9$ ).

The Old Way: Previous proofs were like trying to solve a maze by walking every single path. It was messy and hard to follow.
The New Way: The authors used a "shortcut." They proved that if you take all the pure classical recipes and mix them together, you get exactly the set of all classical recipes for these sizes. They didn't just confirm it; they did it with a fresh, cleaner method that makes the logic much easier to see.

2. The "Directed Graph" Map (The Complex Case)

This is the paper's biggest innovation. What if the puzzle size is complicated, like $d = 12$ (which is $2 \times 2 \times 3$ )? The old "Yes" rule doesn't always work here.

To handle this, the authors invented a Directed Graph.

The Analogy: Imagine a video game map.
- Vertices (The Dots): Each dot represents a specific group of "pure" classical recipes.
- Edges (The Arrows): The arrows show how you can move from one group of recipes to another by changing the size of the puzzle slightly (like swapping a factor of 2 for a factor of 3).
- The Path: You start at the "Start Dot" (where the puzzle is all one type of ingredient) and walk along the arrows to the "End Dot" (where the puzzle is the other type).

The Discovery:
The authors proved a beautiful rule: If you walk along any valid path on this map, the "cloud" of recipes you can make by mixing the groups on that path is exactly the set of classical recipes allowed in that specific section of the map.

It's like saying: "If you only walk down the Red Path on the map, you can only build houses using the bricks found on the Red Path. You can't use bricks from the Blue Path."

Why This Matters

It Unifies Everything: This new "Graph Map" method is a master key. It explains the simple cases (prime powers) and the complex cases (like the $d=12$ example) all in one go. It even proves that previous, smaller discoveries were just special cases of this bigger map.
It Solves the "Counter-Example" Mystery: It explains why the simple "Yes" rule fails for some numbers (like 6). It's not that the rule is broken; it's that you have to look at the specific "paths" on the map to see which recipes are allowed.
Future Tech: Understanding exactly which quantum states are "classical" and which are "quantum" is crucial for building better quantum computers. If we know the boundaries, we know where to look for quantum advantages.

Summary in One Sentence

The authors created a mathematical map (a directed graph) that acts like a GPS for quantum states, showing us exactly how to mix simple "pure" states to build complex "classical" states, proving that the rules of the road depend entirely on the specific path you choose to travel.

1. Problem Statement

The paper addresses the structural characterization of Kirkwood-Dirac (KD) classical states.

Context: The KD quasiprobability distribution is a tool used to describe quantum states via two orthonormal bases, $A = \{|a_i\rangle\}$ and $B = \{|b_j\rangle\}$ . A state $\rho$ is KD-classical if its KD distribution $Q_{ij}(\rho) = \langle a_i|\rho|b_j\rangle\langle b_j|a_i\rangle$ is non-negative for all $i, j$ , effectively forming a valid classical probability distribution. Otherwise, it is KD-nonclassical.
Specific Challenge: The authors focus on the case where the transition matrix $U$ between the two bases is a Discrete Fourier Transform (DFT) matrix ( $U_{ij} = \frac{1}{\sqrt{d}}\omega^{ij}_d$ ).
Open Question: While it was known that for prime power dimensions ( $d=p^r$ $d = p^{r}$ ), the set of KD-classical states ( $KD^+_{A,B}$ $K D_{A, B}^{+}$ ) is the convex hull of KD-classical pure states, this property does not hold for arbitrary dimensions (e.g., $d=6$ $d = 6$ ). Previous work (e.g., Xu [41], Yang et al. [42]) established this for specific cases like $d=p^2$ $d = p^{2}$ or $d=p^r$ $d = p^{r}$ , but a unified structural description for arbitrary dimensions $d$ $d$ was lacking. The paper aims to:
1. Provide an alternative proof for the $d=p^r$ case.
2. Develop a general framework to characterize KD-classical states for any dimension $d$ using a novel graph-theoretic approach.

2. Methodology

The authors employ a combination of algebraic analysis (using $p$ -adic expansions and finite group representations) and graph theory.

Algebraic Framework:
- They utilize the $p$ -adic expansion of integers to decompose the dimension $d$ based on its prime factorization ( $d = p_1^{r_1} \dots p_n^{r_n}$ ).
- They analyze KD-classical pure states, which are defined as superpositions of basis states with specific phase factors determined by the DFT matrix. These states are parameterized by factorizations $d=xy$.
- They utilize Lemma 2 (and its generalization, Lemma 5) to establish linear relationships between sets of KD-classical pure states corresponding to different factorizations of $d$ . Specifically, they show that summing over certain indices in one set of states yields a linear combination of states in a related set.
Graph-Theoretic Construction:
- Directed Graph $G(x_0)$ : For a dimension $d$ $d$ and an initial factorization $d = x_0 y_0$ $d = x_{0} y_{0}$ (with $\gcd(x_0, y_0)=1$ $g cd (x_{0}, y_{0}) = 1$ ), the authors construct a directed graph.
  - Vertices ( $V$ ): Each vertex $v_x$ represents a set of KD-classical pure states associated with a specific factorization $d = xy$.
  - Edges ( $E$ ): A directed edge exists between $v_x$ and $v_{ex}$ if the factorizations differ by exactly one prime factor $p_i$ of $d$ . The direction is determined by whether the prime factor moves from the "x" part to the "y" part or vice versa.
- Paths: A path from the start vertex $v_{x_0}$ to the end vertex $v_{y_0}$ represents a sequence of factorizations where prime factors are gradually transferred between the two sets.
- Linear Space Intersection: The authors define the intersection of the KD-classical state set with the real linear space spanned by the pure states along a specific path.

3. Key Contributions

Alternative Proof for Prime Power Dimensions ( $d=p^r$ ):
- The authors provide a new analytical proof (Theorem 1) confirming that for $d=p^r$ , the set of KD-classical states is exactly the convex hull of KD-classical pure states:
  $KD^+_{A,B} = \text{conv}(\text{pure}(KD^+_{A,B}))$
- This validates a recent result by De Bièvre et al. using a different method involving $p$ -adic expansions and coefficient manipulation.
General Structural Theorem via Directed Graphs (Theorem 2):
- For arbitrary dimensions $d$ , the authors prove that the convex hull of KD-classical pure states along any path from the start vertex to the end vertex in the constructed directed graph is exactly the intersection of the KD-classical state set and the linear space spanned by those specific pure states:
  $KD^+_{A,B} \cap \text{span}_{\mathbb{R}}(v_{\text{path}}) = \text{conv}(v_{\text{path}})$
- This result unifies previous scattered findings. It implies that while the global convex hull property may fail for arbitrary $d$ , it holds locally within the linear subspaces defined by specific paths in the graph.
Unification of Existing Results:
- The paper demonstrates that Theorem 1 (for $d=p^r$ ) and Theorem 2 from Yang et al. [42] (for $d=pq$) are special cases of their general Theorem 2.
- For $d=p^r$ , the graph has a single path covering all pure states.
- For $d=pq$, the graph allows for multiple paths, recovering the specific subset results found in prior literature.

4. Key Results

Lemma 1 & 2: Established orthogonality and summation properties of KD-classical pure states under DFT transitions.
Theorem 1: Confirmed $KD^+_{A,B} = \text{conv}(\text{pure}(KD^+_{A,B}))$ for $d=p^r$ .
Theorem 2: The main result. For any path in the graph $G(x_0)$ , the intersection of the KD-classical set and the span of path-associated pure states is the convex hull of those pure states.
Counter-example Context: The paper acknowledges that for $d=6$ (not a prime power), the global equality $KD^+_{A,B} = \text{conv}(\text{pure}(KD^+_{A,B}))$ fails. The graph approach explains why and provides the correct structural description for such dimensions by restricting the analysis to specific paths.

5. Significance

Theoretical Unification: The work provides a comprehensive theoretical framework that subsumes previous isolated results regarding KD-classicality. It resolves the ambiguity about the structure of KD-classical states in arbitrary dimensions by introducing a topological (graph-based) perspective.
New Analytical Tool: The introduction of directed graphs to represent the algebraic relationships between KD-classical pure states is a novel methodological contribution. It transforms a complex algebraic problem into a structural path-finding problem.
Implications for Quantum Advantage: Since KD-nonclassicality is linked to quantum advantages in tasks like metrology, chaos, and weak values, understanding the precise boundaries of KD-classical states is crucial. This paper clarifies the geometric structure of the "classical" region in the state space, which is essential for identifying and quantifying quantum resources.
Future Directions: The authors suggest that this graph-based approach can be extended to investigate KD-classical states associated with non-DFT transition matrices, potentially opening new avenues for studying quantum non-classicality in broader contexts.

In summary, this paper advances the understanding of KD-classicality by proving that while the global convex hull property is dimension-dependent, a robust local convex hull property exists along specific paths in a graph constructed from the prime factorization of the Hilbert space dimension.