⚛️ quantum physics

Generalized Inverses of Quantum Channels: a categorical perspective

This paper employs a categorical perspective to prove that the Drazin inverse of any quantum channel is trace-preserving, and further demonstrates that for unital channels, both the Drazin and Moore-Penrose inverses preserve unitality, thereby facilitating their application in quantum error mitigation.

Original authors: Robin Cockett, Jean-Simon Pacaud Lemay, Priyaa Varshinee Srinivasan

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Robin Cockett, Jean-Simon Pacaud Lemay, Priyaa Varshinee Srinivasan

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: Fixing a Broken Quantum Machine

Imagine you have a very delicate, high-tech machine (a Quantum Channel) that processes information. In the real world, this machine is noisy. It takes an input, does its job, but accidentally adds some static or "noise" to the output.

Sometimes, this machine is broken in a way that you can simply run it backward to fix it. But often, the noise is irreversible. You can't just hit "rewind."

This is where Generalized Inverses come in. Think of a generalized inverse not as a perfect "undo" button, but as a smart repair manual. It tells you how to process the noisy output to get as close as possible to the original clean input.

In the world of quantum computing, there are two main types of repair manuals:

The Moore-Penrose Inverse: A very popular, standard repair manual.
The Drazin Inverse: A specialized, slightly more complex repair manual.

The Problem: The Repair Manual Might Break the Rules

In quantum physics, there are strict rules for how a machine can operate:

Trace Preserving (TP): The machine must conserve "probability." If you put in 100% of a signal, you must get 100% of a signal out (even if it's noisy). You can't lose the signal into the void.
Completely Positive (CP): The machine must behave physically. It can't create "negative probabilities" or impossible states.

Here is the catch: While the Drazin Inverse always follows the "conservation of signal" rule (TP), the Moore-Penrose Inverse often breaks it. It might try to fix the noise but accidentally make the total signal disappear or explode.

Furthermore, neither of these "repair manuals" is guaranteed to be a physical machine itself (they often fail the CP rule). But for computer simulations and error correction, we really want our repair manuals to at least conserve the signal (TP).

The Solution: A New Way of Looking at the Problem

The authors of this paper decided to stop looking at these machines as just lists of numbers (linear algebra) and started looking at them as abstract shapes and connections (Category Theory).

Think of Category Theory as a universal language of "flow." Instead of calculating specific numbers, they looked at the structure of how information flows through these machines.

By using this high-level "flow" language, they achieved three major breakthroughs:

1. A Simpler Proof for the Drazin Inverse

Previously, proving that the Drazin Inverse always conserves the signal (TP) required pages of heavy, messy math (like trying to fix a car engine by taking it apart bolt-by-bolt).

The Paper's Trick: They used a "categorical" view, which is like looking at the car's blueprint. They showed that because of the way the Drazin Inverse is built, it must conserve the signal, just by the shape of the blueprint itself. It's a much shorter, cleaner proof.
Bonus: They also proved that if the original machine was "Unital" (a specific type of balanced machine), the Drazin repair manual is also balanced.

2. Saving the Moore-Penrose Inverse

The Moore-Penrose Inverse is usually the "go-to" tool, but it was known to be unreliable for quantum channels because it often fails to conserve the signal.

The Paper's Discovery: They found a special condition. If the original quantum machine is Unital (balanced in a specific way), then its Moore-Penrose repair manual will conserve the signal!
Why it matters: This opens the door to using the more popular Moore-Penrose tool for a huge class of quantum machines (called "Mixed Unitary Channels") that are central to cryptography and error correction.

3. Building Better Repair Manuals from Scratch

The paper also looked at how to build complex machines out of simple "pure" parts.

The Analogy: Imagine building a complex repair manual by stacking simple, perfect Lego blocks.
The Result: They showed that if you combine these simple blocks in a specific way (where they don't interfere with each other), the resulting complex machine and its repair manual will both follow all the physical rules (TP and CP). This gives engineers a recipe for building quantum channels that are easy to fix.

The "Category Theory" Magic

Why use this abstract "Category Theory" approach?
Imagine you are trying to explain how a river flows.

Linear Algebra is like measuring the water speed at every single drop. It's accurate but exhausting.
Category Theory is like looking at the map of the riverbed. You don't need to measure every drop to know that water flows downhill and conserves its volume.

The authors used the "map" approach to show that the properties of these quantum repair manuals are baked into the very structure of the universe they operate in, not just a result of complex calculations.

Summary for the Everyday Reader

The Goal: We need tools to fix noisy quantum computers.
The Issue: The best tools often break the fundamental rule of "conserving energy/signal."
The Breakthrough: By viewing these tools through the lens of abstract mathematics (Category Theory), the authors proved:
1. The Drazin Inverse is always a safe tool for conserving signals.
2. The Moore-Penrose Inverse is also safe, provided the machine being fixed is of a specific "balanced" type.
The Impact: This gives quantum engineers new, reliable mathematical tools to build better error-correction systems, helping us move closer to reliable, large-scale quantum computers.

1. Problem Statement

In the era of Noisy Intermediate-Scale Quantum (NISQ) computing, Quantum Error Mitigation (QEM) is crucial for reliable computation. Unlike Quantum Error Correction, which aims to eliminate noise via reversible inverse channels, QEM often relies on generalized inverses of quantum channels to post-process measurement results when noise is non-reversible.

A quantum channel is defined as a Completely Positive (CP) and Trace-Preserving (TP) linear map. While generalized inverses (such as the Moore-Penrose and Drazin inverses) are useful for QEM, they face a significant physical constraint:

The CP Problem: Generalized inverses of quantum channels are generally not Completely Positive, meaning they cannot be physically implemented as quantum operations.
The TP Problem: For simulation stability and physical consistency, it is highly desirable for the generalized inverse to remain Trace-Preserving (TP).
- The Drazin inverse of a TP map is known to be TP, but its Moore-Penrose (MP) inverse is not generally TP.
- Existing proofs for the TP property of the Drazin inverse rely on heavy linear algebra (e.g., Jordan–Chevalley decomposition).

The Core Question: Can we provide a more streamlined, structural proof that generalized inverses preserve the TP (and Unital) properties of quantum channels? Furthermore, under what conditions do these inverses preserve the TP property for the Moore-Penrose case, and can we identify classes of channels where the generalized inverse remains a valid quantum channel (i.e., CP, TP, and Unital)?

2. Methodology

The authors adopt a categorical perspective, utilizing Dagger Compact Closed Categories (†KCC) to abstract quantum channels. This framework allows for the study of quantum channels without relying on specific matrix representations or coordinate-based linear algebra.

Framework: The paper operates within †KCC, where objects represent Hilbert spaces and maps represent linear transformations.
- CP Maps: Defined via the "Choi-Jamiołkowski" isomorphism using the internal hom and trace.
- TP Maps: Defined via the preservation of the internal trace operator ( $\text{tr}_A$ ).
- Unital (U) Maps: Defined via the preservation of the unit map ( $u_A$ ).
Generalized Inverses: The authors utilize categorical definitions of:
- Drazin Inverse ( $f^D$ ): Defined for endomaps satisfying specific algebraic conditions (index $k$ , commutativity, etc.).
- †-Drazin Inverse ( $f^\partial$ ): A generalization for arbitrary maps in dagger categories.
- Moore-Penrose Inverse ( $f^\circ$ ): A specific case of the †-Drazin inverse satisfying additional adjoint conditions.
Key Technique: The authors leverage commuting diagrams and the algebraic properties of generalized inverses (specifically Proposition 3.4 from prior literature) to prove preservation of TP and U properties. They avoid explicit matrix calculations, instead using the structural properties of the category (e.g., naturality of trace, properties of the dagger functor).

3. Key Contributions and Results

A. Streamlined Proof for Drazin Inverses (TP and Unital)

Result: The authors prove that if a quantum channel $\Phi$ is Trace-Preserving (TP), its Drazin inverse $\Phi^D$ is also TP.
Method: They construct a simple commuting triangle involving the trace map ( $\text{tr}$ ) and the identity map. Since the identity is trivially Drazin and the trace is a natural transformation, the commutativity of the diagram for $\Phi$ implies the commutativity for $\Phi^D$ .
Extension: They similarly prove that if $\Phi$ is Unital (U), then $\Phi^D$ is also U.
Significance: This provides a significantly shorter and more conceptual proof than the heavy linear algebra used in previous literature (e.g., [5, Thm 2]).

B. Dagger Drazin and Moore-Penrose Inverses

Limitation of Drazin: Standard Drazin inverses are only defined for endomaps ( $A \to A$ ).
Solution: The authors utilize †-Drazin inverses (defined for $A \to B$ ) and Moore-Penrose (MP) inverses.
Key Finding: While the MP inverse of a TP map is not generally TP, the authors prove a strong equivalence for Unital channels:
- Theorem: A map $\Phi$ is TP and Unital if and only if its Moore-Penrose inverse $\Phi^\circ$ is TP and Unital.
Implication: This establishes that for Unital Quantum Channels (a broad class including mixed unitary channels), the Moore-Penrose inverse is a valid candidate for error mitigation as it preserves both TP and Unital properties.

C. Preservation of Complete Positivity (CP)

Observation: Generalized inverses generally do not preserve Complete Positivity (CP). The paper provides a counter-example using depolarizing channels where the Drazin/MP inverse exists but fails to be CP.
Positive Case (Pure/Mixed Channels): The authors identify specific classes of channels where the generalized inverse does remain CP (and thus a valid quantum channel):
- Pure Channels: Maps of the form $\Phi = [f^\dagger, f]$ .
- Mixed Pure Channels: Sums of pure channels $\Phi = \sum [f_i^\dagger, f_i]$ under orthogonality conditions ( $f_i f_j^\dagger = 0$ for $i \neq j$ ).
Result: For these specific mixed pure channels, the generalized inverse (Drazin, †-Drazin, or MP) is not only TP and Unital but also CP. This means the inverse is a physically implementable quantum channel.

4. Significance and Applications

Theoretical Unification: The paper bridges the gap between categorical quantum foundations (dagger compact closed categories) and the algebraic theory of generalized inverses. It demonstrates that categorical abstractions can yield simpler, more robust proofs for properties of quantum channels.
Quantum Error Mitigation (QEM):
- The results validate the use of Drazin inverses for general TP channels in QEM protocols.
- Crucially, the paper opens the door for using Moore-Penrose inverses in QEM for Unital Quantum Channels (e.g., mixed unitary channels), a class central to noise modeling in quantum cryptography and decoherence. Previously, the lack of TP preservation for MP inverses limited their applicability.
New Class of Reversible Channels: By identifying "Mixed Pure Channels" where the generalized inverse remains a valid quantum channel (CP, TP, U), the paper suggests new protocols where error mitigation can be performed via a physically realizable inverse operation, rather than just a mathematical post-processing step.
Future Directions: The work highlights the potential of †-TSMCC (Dagger Traced Symmetric Monoidal Closed Categories) as a setting for further study, though current examples are limited to †KCC.

Conclusion

Cockett, Lemay, and Srinivasan successfully demonstrate that a categorical approach simplifies the analysis of generalized inverses in quantum information. They prove that Drazin inverses preserve TP and Unital properties universally, and that Moore-Penrose inverses preserve these properties specifically for Unital channels. Furthermore, they identify a specific class of "Mixed Pure" channels where the generalized inverse is a fully valid quantum channel, offering new theoretical tools for designing robust quantum error mitigation strategies.