🔢 mathematics

Existence of Kraus decomposition in infinite dimension via strongly-convergent direct process tomography

This paper presents a constructive algorithm for obtaining a Kraus decomposition of completely positive operators on infinite-dimensional Hilbert spaces by iteratively generating operators with increasing zero entries, thereby ensuring strong-operator convergence and improving upon standard nonconstructive proofs through practical process tomography.

Original authors: Paul E. Lammert

Published 2026-08-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Paul E. Lammert

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

Imagine the universe as a giant, invisible stage where particles perform a delicate dance. Sometimes, these dancers are perfect and predictable, but often, they bump into the audience, trip over their own feet, or get distracted by the lights. In the world of quantum physics, this "messiness" is called an "open system," and the rules that describe how a quantum state changes when it gets messy are called "completely positive maps." Think of these maps as the rulebook for how a quantum coin flips, spins, or lands when the wind blows.

To understand these rulebooks, scientists use a special toolkit called "Kraus decomposition." Imagine trying to explain a complex magic trick. Instead of describing the whole confusing routine at once, you break it down into a list of simple, individual moves. Each move is a "Kraus operator." If you add up all these simple moves, you get the full, messy magic trick back again. For a long time, scientists knew these lists existed for small, simple systems (like a single atom), but when they tried to apply this to huge, infinite systems (like a whole field of atoms), the math got scary. The old proofs were like saying, "A list exists, trust us," without showing you how to actually write it down. They were non-constructive, meaning they proved the answer was there but gave no instructions on how to find it.

This paper, titled "Existence of Kraus decomposition in infinite dimension via strongly-convergent direct process tomography," steps in to fix that. The author, Paul E. Lammert, presents a clever, step-by-step recipe (an algorithm) that actually builds this list of moves, one by one, even for infinite systems. They don't just say the list exists; they show you exactly how to generate it, proving that as you add more and more moves to your list, the sum gets closer and closer to the true, messy reality. It's like finally getting the instruction manual for the infinite magic trick, complete with a guarantee that if you follow the steps, you'll get the right result.

The Infinite Puzzle and the "Zero" Strategy

In the quantum world, scientists often try to figure out what a machine is doing by poking it and watching what comes out. This is called "process tomography." Usually, you measure the machine, get a bunch of numbers, and then try to reverse-engineer the rulebook (the Kraus operators) from those numbers. The problem is, when the machine is infinitely complex, the math gets stuck. The old way of thinking suggested that for infinite systems, you might need a continuous "integral" (a smooth, flowing sum) rather than a list of distinct steps. It felt like you couldn't count your way to infinity.

The author of this paper says, "Not so fast!" They propose a method that is both "constructive" (it builds the answer) and "elementary" (it doesn't need super-complex, abstract math to work). Their big idea is to treat the infinite problem like a giant puzzle where you fill in the blanks one by one.

Here is how their algorithm works, using a playful analogy:

Imagine you have a giant, infinite grid of light switches. Each switch represents a possible interaction between the quantum system and its environment. Some switches are "on" (active), and some are "off" (zero). Your goal is to figure out exactly which switches are on to describe the system's behavior.

The author's algorithm starts with the whole messy grid. It picks a specific pair of coordinates—a specific "row" and "column" (which they call a pair of vectors, $h$ and $k$ )—and asks, "Is there any activity here?"

The Check: If the activity is zero, great! They move on.
The Extraction: If there is activity, they calculate a specific "Kraus operator" (a simple move) that explains exactly that piece of activity.
The Subtraction: They subtract this new move from the original messy grid.
The Magic of Zeros: Here is the clever part. Because of the way they calculate the move, the specific spot they just looked at is now guaranteed to be zero in the remaining grid. It's like they just turned off that specific light switch and locked it in the "off" position.

They repeat this process over and over, picking new pairs of coordinates in a specific order. With every step, they generate a new Kraus operator and leave behind a "remainder" grid that has one more guaranteed zero spot than before.

Why This Matters: The "Stream" of Answers

The beauty of this method isn't just that it finds the answer; it's how it finds it. The author proves that this stream of generated operators converges. In plain English, this means that if you stop the algorithm after 10 steps, you have a very good approximation of the system restricted to a small part of the universe. If you stop after 1,000 steps, you have a better approximation of a larger part. If you let it run forever, the sum of all these steps perfectly reconstructs the original infinite system.

The paper explicitly rules out the idea that you must use integrals or non-constructive proofs for infinite dimensions. They show that a simple, step-by-step sum is sufficient. They also argue against the notion that finding these decompositions is impossible or purely theoretical. By fusing the abstract math with "direct process tomography" (a practical way of measuring systems), they turn a philosophical question ("Does it exist?") into a practical engineering task ("Here is how you build it").

The author is very confident in their results. They don't just simulate this on a computer; they provide a rigorous mathematical proof that the algorithm works. They prove two critical things:

The operators they generate are "bounded," meaning they don't blow up to infinity and break the math.
The remainder (the part of the system you haven't explained yet) shrinks down to nothing as you add more steps.

A Practical Takeaway

Why should a curious teenager care? Because this paper bridges the gap between the "impossible" and the "doable." In quantum computing and quantum communication, we are moving from small, simple experiments to massive, complex networks. To design these networks, we need to understand how they lose information (noise).

The author shows that we can treat these infinite, noisy systems as a series of manageable chunks. The algorithm they present acts like a "progress bar" for understanding quantum noise. You can stop the process at any time, and you'll have a valid, working description of the system for a specific size. This is incredibly useful for engineers who might not need the entire infinite solution, just a very good approximation for a large-but-finite system.

In short, this paper takes a scary, infinite math problem and solves it with a simple, repetitive recipe. It proves that even in the infinite dark of quantum mechanics, you can find your way by turning on the lights one by one, knowing that every light you turn on brings you closer to the whole picture. The author has provided the map, and they've proven that if you follow the path, you will get there.

Technical Summary: Existence of Kraus Decomposition in Infinite Dimension via Strongly-Convergent Direct Process Tomography

Problem Statement
The paper addresses the construction of Kraus decompositions (operator-sum representations) for completely positive (CP) operators acting on separable, countably-infinite-dimensional Hilbert spaces. While the existence of such decompositions is a fundamental result in quantum information and open quantum systems, standard proofs for infinite dimensions (e.g., Kraus [10,11]) rely on sophisticated $C^*$ -algebra representation theory and are nonconstructive. They fail to provide a method for actually finding the decomposition or approximating it. Furthermore, existing finite-dimensional proofs do not naturally generalize to the infinite case, where one might expect integral representations rather than sums. The specific challenge is to provide a constructive algorithm that generates a sequence of Kraus operators $\{A_i\}$ such that the sum $\sum \Theta(A_i)$ converges to the target CP map $\Lambda$ in the strong-operator topology (SOT).

Methodology
The author proposes a "direct process tomography" algorithm that constructs Kraus operators one-by-one. The method relies on the following components:

Ground Matrix Elements and Kernel Relations: The algorithm operates on the "ground matrix elements" of the CP map $\Lambda$ , defined as $\langle k'k | \Lambda \cdot h'h \rangle$ . To track progress, the author defines a "kernel relation" $K_\Lambda$ , a set of pairs $(k, h)$ for which the corresponding matrix elements vanish.
The Reduction Step (Proposition 1): The core of the algorithm is a reduction step. Given a CP map $\Lambda$ and a pair $(k, h)$ not in its kernel relation, the author constructs a unique (up to a phase) bounded operator $A$ such that $\Lambda' = \Lambda - \Theta(A)$ is still CP and $(k, h)$ becomes an element of the kernel relation of $\Lambda'$ . The operator $A$ is explicitly defined by:
$\langle k' | A h' \rangle = \frac{\langle k'k | \Lambda \cdot h'h \rangle}{\sqrt{\langle \Pi(k) | \Lambda \cdot \Pi(h) \rangle}}$
This step effectively "removes" the contribution of the specific matrix element associated with $(k, h)$ while preserving the CP property.
Iterative Process: The algorithm enumerates pairs from orthonormal bases of the input and output Hilbert spaces, $(k_n, h_n)$ . At each step $n$ , it applies the reduction step to the remainder map $\Lambda_{n-1}$ to generate $A_n$ and the new remainder $\Lambda_n = \Lambda_{n-1} - \Theta(A_n)$ .
Matrix View and Zero Propagation: The author visualizes the process using a "matrix of matrices" perspective. As $n$ increases, the operators $A_n$ and the remainders $\Lambda_n$ acquire guaranteed zero entries in their matrix representations. Specifically, for a given $n$ , the entries corresponding to pairs $(k_m, h_m)$ with $m \leq n$ are guaranteed to be zero in the remainder $\Lambda_n$ .

Key Contributions and Results

Constructive Proof: The paper provides an elementary, constructive proof for the existence of Kraus decompositions in infinite dimensions. Unlike previous proofs, this approach explicitly generates the operators.
Strong-Operator Topology Convergence: The author proves that the sequence of partial sums $\sum_{i=1}^n \Theta(A_i)$ converges to $\Lambda$ in the strong-operator topology. This is equivalent to proving that the remainder maps $\Lambda_n$ converge to zero in the SOT.
Boundedness and CP Preservation: The proof establishes that the generated operators $A_n$ are bounded and that every intermediate remainder $\Lambda_n$ remains a completely positive map.
Uniform Bounds: The algorithm maintains uniform bounds on the norms of the generated operators and remainders, specifically $\|A_n\|^2, \|\Lambda_n\|_{1,1} \leq \|\Lambda\|_{1,1}$ .
Finite-Dimensional Approximations: The algorithm naturally provides exact Kraus decompositions for the restrictions (projections) of $\Lambda$ onto increasingly larger finite-dimensional subspaces. The intermediate results $\Lambda = \sum_{i=1}^N \Theta(A_i) + \Lambda_N$ contain all information about the map's restriction to these subspaces.

Significance and Claims
The paper claims that this approach improves upon standard proofs by fusing abstract existence questions with practical process tomography. The significance lies in:

Elementary Nature: The proof avoids advanced $C^*$ -algebra machinery, relying instead on elementary properties of positive operators and the quasi-Cauchy-Schwarz inequality.
Constructivity: It offers a concrete procedure to find the decomposition, addressing the nonconstructive nature of prior infinite-dimensional proofs.
Practical Utility: The method is described as practical for application, particularly for large-but-finite dimensional systems. The "stream of outputs" provides a coherent family of decompositions for restrictions of the target map, allowing for systematic approximation.
Theoretical Completeness: By proving the SOT convergence of the remainders, the work rigorously establishes the existence of the infinite sum representation for separable Hilbert spaces, confirming that integrals are not generally required even in infinite dimensions.

The author does not propose new experimental setups or specific future applications beyond the utility of the algorithm for process tomography and the theoretical clarification of Kraus decompositions in infinite dimensions.

The Infinite Puzzle and the "Zero" Strategy

Why This Matters: The "Stream" of Answers

A Practical Takeaway

Technical Summary: Existence of Kraus Decomposition in Infinite Dimension via Strongly-Convergent Direct Process Tomography

More like this