In classical physics, you can measure every property of a system simultaneously without interference. In quantum mechanics, however, certain measurements are fundamentally incompatible: measuring one inevitably disturbs the other, as famously exemplified by position and momentum.
For decades, measurement incompatibility was treated as a binary property—measurements were either compatible (jointly measurable) or they were not. However, modern quantum information theory views incompatibility not just as a philosophical constraint, but as a valuable operational resource for tasks like quantum cryptography and Einstein-Podolsky-Rosen (EPR) steering. To utilize this resource, we must move beyond a simple "yes or no" and learn to precisely quantify how incompatible a set of measurements actually is.
The Intuition of Noise Robustness
How do we measure the degree of a purely quantum property? One of the most intuitive approaches is to ask: How much noise is required to destroy it?
Imagine a pair of quantum measurements. If they are only mildly incompatible, even a small amount of environmental noise (or unsharpness) will blur them enough that they become jointly measurable via a single, classical apparatus. If they are deeply incompatible, they will withstand a massive amount of noise before their quantum friction disappears.
This intuition forms the basis of "incompatibility robustness." In Quantifying Measurement Incompatibility of Mutually Unbiased Bases, Skrzypczyk et al. applied this metric to Mutually Unbiased Bases (MUBs)—measurement configurations that are traditionally considered "maximally incompatible." By rigorously calculating their noise robustness, the authors demonstrated that not all MUBs are created equal; different configurations possess different operational strengths when demonstrating quantum steering, proving that qualitative maximality does not imply quantitative equivalence.
Incompatible Measurements + White Noise ──> Jointly Measurable
└───────────────┬───────────────┘
Amount of noise tolerated
= Incompatibility Robustness
However, quantifying incompatibility is rarely straightforward. As Farkas et al. demonstrated in Incompatibility robustness of quantum measurements: a unified framework, the definition of which measurements are the "most incompatible" is highly relative. It depends entirely on the mathematical structure of the noise model and the specific metric chosen. The authors established a unified framework to show that many existing measures failed to satisfy basic theoretical axioms, highlighting the need for metrics that remain mathematically consistent across different quantum tasks.
Dimensionality and the Subspace Problem
Incompatibility is intimately tied to the size of the sandbox the quantum system plays in. High-dimensional systems offer more "room" for measurements to misalign, creating complex landscapes of incompatibility.
To understand this deeply, we can look at the problem through two lenses: compression and restriction.
- Compression ($n$-simulability): Can the incompatibility of a high-dimensional system be simulated by a smaller one? In Simulability of High-Dimensional Quantum Measurements, Ioannou et al. introduced the concept of "$n$-simulability." Instead of asking if measurements can be performed simultaneously, they asked whether a high-dimensional measurement set can be perfectly mimicked by routing the quantum state through an $n$-dimensional bottleneck. This provides a way to quantify incompatibility in terms of the physical dimensions required to manifest it.
- Restriction (Subspaces): Conversely, if we take a highly incompatible high-dimensional system and restrict our focus to a smaller subspace, does the incompatibility survive? Uola et al. explored this geometric puzzle in Quantum measurement incompatibility in subspaces, probing whether subspace-dependent incompatibility preserves properties like convexity and developing "witnesses" to detect these localized quantum frictions in practical scenarios.
Measuring Incompatibility in the Lab
While robustness and simulability provide excellent theoretical frameworks, calculating them often requires solving demanding optimization problems that assume you already know everything about the measurement devices. In practice, executing full tomography on an unknown quantum device is notoriously difficult. How do we measure incompatibility experimentally without full data access?
One elegant solution leverages the concept of Mutual Eigenspace Disturbance (MED). If two observables are compatible, measuring the first will not alter the probabilities of the second's outcomes. If they are incompatible, the first measurement collapses the state, violently disrupting the eigenspaces of the second.
In Measuring Incompatibility and Clustering Quantum Observables with a Quantum Switch, Li et al. introduced a method to estimate this disturbance directly using a quantum switch. A quantum switch places the order of two operations into a quantum superposition (i.e., measuring $A$ then $B$, superposed with measuring $B$ then $A$).
Control Qubit ───► Superposition of Orders: (|A then B⟩ + |B then A⟩)
│
▼
Interference Pattern ───► Directly yields MED (Incompatibility)
By observing how the control qubit of the switch loses coherence, researchers can directly read out the MED without ever looking at the actual outcomes of measurements $A$ or $B$. This bypasses experimental complexity, allowing unsupervised machine learning algorithms to cluster completely unknown observables based purely on their measured incompatibility.
A parallel challenge arises in quantum machine learning, where extracting classical information from quantum states introduces noise. In New Bounds on Quantum Sample Complexity of Measurement Classes, Heidari et al. designed a specialized norm to quantify the incompatibility of measurement classes used in creating "classical shadows" of quantum states. By measuring this incompatibility mathematically, they proved that learning from quantum data does not inherently require exponentially more samples than classical learning, bounding the physical limitations of information extraction.
The Inverse Problem: Mapping the Compatible Space
To truly master incompatibility as a resource, we must also understand its inverse: given a specific reference measurement, what is the complete set of all other measurements that are compatible with it?
This characterization has historically been an open challenge, but Pellonpää et al. provided a constructive framework in Naimark dilations of qubit POVMs and joint measurements. Their approach relies on Naimark's dilation theorem, which states that any generalized quantum measurement (a Positive Operator-Valued Measure, or POVM) can be viewed as a standard, sharp projection measurement if you "dilate" the system by coupling it to a larger, auxiliary Hilbert space.
The authors showed that when you lift a reference measurement into this larger space, any measurement compatible with it is forced to adopt a highly constrained, block-diagonal structure. By mapping out this geometric structure in the dilated space, researchers can systematically construct the entire family of compatible partners for a given measurement, closing the loop on how we bound, measure, and utilize quantum incompatibility.
Key Foundational Work
Incompatibility robustness of quantum measurements: a unified framework (2019) — contributed 6 concept elements to this direction.
Quantifying Measurement Incompatibility of Mutually Unbiased Bases (2019) — contributed 2 concept elements to this direction.
Quantum measurement incompatibility in subspaces (2021) — contributed 1 concept elements to this direction.