Magic Monotones in Quantum Computation

• Magic monotones are convex, monotonic functionals on quantum states that vanish for stabilizer mixtures, serving as precise measures of non-stabilizerness in quantum computation.
• They integrate geometric, entropic, operational, and algebraic methods—such as facet analysis and Rényi entropies—to provide rigorous conversion constraints and simulation bounds.
• While essential for state certification and circuit simulation, magic monotones are often super-exponentially hard to compute, posing significant practical challenges.

A magic monotone is any convex, monotonic functional on quantum states (or channels) that vanishes exactly for mixtures of stabilizer states—quantifying non-stabilizerness (“magic”) as a computational resource for universal quantum computation. Magic monotones play a central role in the resource theory of magic, underpinning rigorous conversion constraints, bounds on classical simulability, and robustness to noise. A rich variety of monotones has emerged, reflecting geometric, entropic, operational, and algebraic aspects of the stabilizer/polytope structure.

1. Formal Axioms and Framework

In the resource theory of magic, the free states are the convex hull of pure stabilizer states (STAB), i.e., density matrices ρ expressible as $\rho = \sum_j p_j |\phi_j\rangle\langle\phi_j|$ with each $|\phi_j\rangle$ a stabilizer vector. Free operations are the completely positive trace-preserving (CPTP) maps that map STAB into itself (stabilizer-preserving operations), generated by Clifford unitaries, computational-basis measurements, preparation of $|0\rangle$, partial traces, and classical randomness (Leone et al., 2024, Ahmadi et al., 2017).

A magic monotone $\mathcal{M}(\rho)$ must satisfy:

• Faithfulness: $\mathcal{M}(\rho) = 0$ if and only if $\rho \in \mathrm{STAB}$.
• Monotonicity: For any free operation $\Lambda$, $\mathcal{M}(\Lambda(\rho)) \leq \mathcal{M}(\rho)$.
• Convexity: $$\mathcal{M}(\sum_i p_i \rho_i) \leq \sum_i p_i \mathcal{M}(\rho_i)$$.
• Strong monotonicity (optional): average nonincrease under selective measurement outcomes. Additional desiderata include invariance under Clifford gates and subadditivity under tensor products (Leone et al., 2024, Leone et al., 25 Feb 2026).

Hardness of Computation

It is now established that, under the Exponential Time Hypothesis (ETH), all faithful and convex magic monotones are super-exponentially hard to compute: deciding STAB membership (faithful detection of magic) for general $n$-qubit states requires $\exp(\Omega(n^2))$ time, a complexity matched by the robustness of magic, which is thus computationally optimal among monotones (Leone et al., 25 Feb 2026).

2. Geometric and Witness-Based Monotones

Minkowski-Functional (Facet) Magic Monotone

Recent advances define magic monotones by direct analysis of the stabilizer polytope's facets in Pauli expectation value space (Warmuz et al., 2024). For any $N$-qubit state, the vector of Pauli expectations $r(\rho)$ satisfies a set of facet inequalities $w(a)\cdot r(\rho) \leq 1$ for integer vectors $a$ (with $a_0 = 0$ for the identity). The magic monotone is

$$M(\rho) = \max_{a \in \mathbb{Z}^{4^N},\,a_0=0} \left[ \frac{a\cdot r(\rho)}{b(a)} - 1 \right],$$

where $b(a) = \max_i a \cdot S_i$ over all stabilizer vertices $S_i$.

• $M(\rho) > 0$ if and only if $\rho$ is outside the stabilizer polytope (i.e., non-stabilizer/magic).
• Computationally, the number of parameters ($4^N-1$) is vastly smaller than the number of stabilizer mixtures ($\sim 2^{N^2/2}$).
• Simpler “witness” variants $W(\rho)$ optimize linear functionals with bounded norm and provide a computationally efficient indicator (Warmuz et al., 2024).

3. Entropic Magic Monotones

Stabilizer Rényi Entropies

For a pure state $|\psi\rangle$, define the Pauli distribution $$\Xi_P(\psi) = \frac{|\langle\psi|P|\psi\rangle|^2}{2^N}$$. The order-$n$ stabilizer entropy is

$$M_n(|\psi\rangle) = \frac{1}{1-n}\log \left[ \sum_P \Xi_P(\psi)^n \right].$$

• $M_n = 0$ if and only if $|\psi\rangle$ is a stabilizer state.
• For $n \geq 2$ (including the “linear” entropy $S_2$, and in particular $L = 1 - \sum_P \Xi_P(\psi)^2$), $M_n$ and $L$ are true monotones under all stabilizer protocols (both deterministic and probabilistic, with strong monotonicity for $L$) (Leone et al., 2024).
• For $0 \leq n < 2$, stabilizer entropies fail weak and strong monotonicity (Haug et al., 2023). The open case $n \geq 2$ for mixed states is resolved via convex roofs (Leone et al., 2024).

These entropies provide both quantitatively tight conversion rates between magic states and are efficiently computable for low-rank or matrix product states (Leone et al., 2024, Haug et al., 25 Apr 2025).

Mixed-State Magic Witnesses

The entropic magic–Rényi witnesses $\mathcal{W}_\alpha$, based on moments of the Pauli spectrum $A_\alpha(\rho)$, provide efficiently measurable, Clifford-invariant tests for magic with numerical lower bounds on robustness and fidelity monotones. They are strictly positive for non-stabilizer states and can be measured using Bell-basis swaps and replica tricks (Haug et al., 25 Apr 2025).

4. Operational, Algebraic, and Channel Magic Monotones

Robustness of Magic (RoM) and Related Quantities

The robustness of magic for states,

$$R(\rho) = \min \left\{ t \geq 0: \rho = (1+t)\sigma_+ - t\sigma_-,\ \sigma_\pm \in \mathrm{STAB} \right\},$$

is the paradigmatic monotone but is computationally hard (super-exponential scaling) (Leone et al., 25 Feb 2026). RoM is monotonic under all stabilizer protocols and determines sample complexity for simulation algorithms (Seddon et al., 2020).

Other related monotones include the stabilizer nullity, stabilizer fidelity, min-relative entropy of magic, and extent (Kalra et al., 6 Mar 2025, Rubboli et al., 2023, Vardhan et al., 13 Mar 2025). The “Barnes-Wall norm” gives a lattice-geometric magic measure for pure states, capturing magic in the integer structure of the Barnes–Wall lattice (Kalra et al., 6 Mar 2025).

Channel-Based Magic Monotones

For quantum channels, magic monotones extend via the Choi isomorphism to:

• Channel Robustness $R_{\mathrm{ch}}(\mathcal{E})$ (Seddon et al., 2019),
• Magic Capacity $C_{\mathrm{magic}}(\mathcal{E})$ (maximal output magic on stabilizer inputs) which quantify the classical simulation cost for noisy operations in circuit models.

Their properties mirror those of state robustness: faithfulness, convexity, monotonicity, and tight operational sample complexity bounds.

Magic in Operator Space

Heisenberg-evolution- or operator-space-based monotones include the operator stabilizer Rényi entropy, the $T$-count, and unitary nullity, which upper-bound local operator entanglement (LOE) and diagnose classical non-simulability for quantum chaotic dynamics (Dowling et al., 30 Jan 2025). These quantifiers are operationally relevant for simulation algorithms and chaoticity detection in quantum circuits.

Magic for Qudits and Wigner Negativity

In odd prime dimensions, magic monotones based on sum-negativity (mana), thauma, and Rényi divergences of the discrete Wigner function give a single-shot statistical mechanics framework, connecting majorization, Lorenz curves, and monotone families (Koukoulekidis et al., 2021). These monotones are Schur-concave, extend to negative quasi-distributions, and lead to strict conversion bounds for distillation and simulation.

5. Completeness, Additivity, and Trade-Offs

Complete Families and Additive Monotones

Certain sets of magic monotones are complete: for example, the two-parameter family $\{M_{\sigma,t}\}$ constructed from conditional min-entropies provides necessary and sufficient criteria for single-shot magic-state convertibility under stabilizer-preserving channels (Ahmadi et al., 2017). These are efficiently computable via semi-definite programs for finite $d$.

Notably, stabilizer fidelity and many relative-entropy-based monotones are multiplicative under tensor products in the single-qubit regime and additive for several classes of multi-qubit and noisy states (Rubboli et al., 2023). Explicit closed-form additivity and conversion bounds are established for T, H, F states, Toffoli, CCZ, and depolarized magic states.

Monotones, Chirality, Discord

Chirality-based monotones, such as the chiral log-distance, serve as lower bounds for stabilizer nullity and stabilizer fidelity and link resource theories of magic, discord, and nonlocal correlations. The chiral log-distance is invariant under local unitaries and vanishes exactly on stabilizer states (Vardhan et al., 13 Mar 2025).

6. Practical Applications and Scaling

• State certification: Efficient witnesses ($\mathcal{W}_\alpha$, facet monotones) allow practical mixed-state magic certification on near-term platforms, including noisy and highly entangled MPS subsystems (Warmuz et al., 2024, Haug et al., 25 Apr 2025).
• Circuit simulation: Magic monotones govern the exponential runtime and precision tradeoffs in stabilizer/non-stabilizer circuit simulation, giving precise scaling for dyadic negativity, extent, and channel robustness (Seddon et al., 2020, Seddon et al., 2019).
• Distillation and conversion rates: Additive and multiplicative properties provide direct, exponentially tight converse and rate bounds for magic-state distillation even in one-shot/probabilistic settings (Rubboli et al., 2023, Leone et al., 2024).
• Noise robustness and cryptography: Magic persists under strong global depolarization, and cryptographic pseudomagic hiding requires an extensive entropy resource (Haug et al., 25 Apr 2025).

7. Limitations and Open Problems

• Intractability in general: Any fully faithful monotone is super-exponential to compute in $n$; exceptions are low-rank, structured, or small-$n$ settings (Leone et al., 25 Feb 2026).
• Classification of all stabilizer polytope facets (large $N$): Facet enumeration becomes challenging but not as intractable as RoM optimization (Warmuz et al., 2024).
• Extensions to qudits, many-body systems, and catalytic processes: Majorization approaches and operator-based monotones point to ongoing generalizations (Koukoulekidis et al., 2021, Dowling et al., 30 Jan 2025).
• Search for monotones that are both strongly monotonic and computationally efficient for general mixed states: No fully satisfactory such monotone is currently known in the high-rank regime (Haug et al., 2023, Leone et al., 2024).
• Physical implementation and experimental readout: While some monotones are experimentally accessible (Pauli sampling, twin measurements), full state tomography or SDP remains challenging for $n > 6$ (Haug et al., 25 Apr 2025).

---

References:

• "A magic monotone for faithful detection of non-stabilizerness in mixed states" (Warmuz et al., 2024).
• "Stabilizer entropies are monotones for magic-state resource theory" (Leone et al., 2024).
• "Efficient witnessing and testing of magic in mixed quantum states" (Haug et al., 25 Apr 2025).
• "Stabilizer entropies and nonstabilizerness monotones" (Haug et al., 2023).
• "Mixed-state additivity properties of magic monotones based on quantum relative entropies for single-qubit states and beyond" (Rubboli et al., 2023).
• "Quantification and manipulation of magic states" (Ahmadi et al., 2017).
• "The unbearable hardness of deciding about magic" (Leone et al., 25 Feb 2026).
• "Stabilizer Ranks, Barnes Wall Lattices and Magic Monotones" (Kalra et al., 6 Mar 2025).
• "Bridging Entanglement and Magic Resources through Operator Space" (Dowling et al., 30 Jan 2025).
• "Resource theory of quantum scrambling" (Garcia et al., 2022).
• "Magic Monotones for Quantum Channels" (Seddon et al., 2019).
• "Constraints on magic state protocols from the statistical mechanics of Wigner negativity" (Koukoulekidis et al., 2021).
• "Quantifying quantum speedups: improved classical simulation from tighter magic monotones" (Seddon et al., 2020).
• "Chirality, magic, and quantum correlations in multipartite quantum states" (Vardhan et al., 13 Mar 2025).