Quantum State Designs via Magic Teleportation

Technical Summary: Quantum State Designs via Magic Teleportation

Problem Statement
The generation of quantum state designs—ensembles that reproduce the first $k$ moments of the Haar-random distribution—is a fundamental task for quantum information applications ranging from tomography to demonstrating quantum advantage. While random unitary circuits can generate these designs, fault-tolerant quantum computing relies heavily on Clifford operations, which alone are insufficient to create high-order designs (they form only a 3-design on the full system but degrade to a 1-design upon projection). The paper addresses the gap in understanding how limited "magic" (non-stabilizer resources) injected into Clifford circuits enables the emergence of $k$-designs within the projected ensemble. Specifically, it investigates how the amount of initial magic and the circuit architecture (depth and locality) control the convergence to random states when a subsystem is measured.

Methodology
The authors analyze a protocol where an initial state $|\psi_0\rangle$ with finite magic is evolved under a Clifford circuit $C_t$ and subsequently subjected to projective Pauli measurements on a subsystem $B$. The resulting ensemble of post-measurement states on the remaining subsystem $A$ is the "projected ensemble."

1. Quantification: The convergence to a $k$-design is quantified using the frame potential $F^{(k)}_E$. The deviation from the Haar-random value is measured by $\Delta^{(k)}_A$.
2. Magic Metric: The resource content of the initial state is quantified via Stabilizer Rényi Entropies ($M_k$), defined as $M_k(|\psi\rangle) = (1-k)^{-1} \log \sum_{s \in P_L} \langle \psi | s | \psi \rangle^{2k}$.
3. Analytical & Numerical Approach:
   • Deep Circuits: The authors derive analytical bounds using the commutant theory of the Clifford group and the replica method. They propose a Magic-Induced Design Ansatz (MIDA) relating the design error to the stabilizer entropy.
   • Shallow Circuits: They simulate local and long-range Clifford circuits to study the timescales of convergence, analyzing how magic propagates from the injection site to the measurement region.
   • State Preparation: Numerical simulations employ protocols to generate random states with fixed stabilizer entropy, including product phase states and random states with fixed Von Neumann magic.

Key Contributions and Results

1. Magic-Induced Design Ansatz (MIDA):
   For deep Clifford circuits, the authors establish that the projected ensemble converges to a $k$-design with an error that decays exponentially with the $k$-th Stabilizer Rényi Entropy ($M_k$) of the initial state.

   • Result: $\Delta^{(k)}_A \propto \exp[(1-k)M_k]$.
   • Validation: This ansatz is supported by rigorous upper bounds derived from Clifford commutant theory and extensive numerical simulations across various initial states (product phase states and random states). The deviation $\Delta^{(k)}_A$ is shown to be upper-bounded by terms involving $2^{-M_2}$, confirming that higher magic content leads to faster convergence to random states.

2. Magic Teleportation Mechanism:
   In shallow circuits, the emergence of designs is governed by the transport of magic. The authors identify a "magic teleportation" phenomenon where non-Clifford resources spread through Clifford scrambling and measurements, effectively violating standard light-cone locality constraints.

   • Local Architectures: When magic is injected far from the measurement subsystem $A$, the convergence timescale $\tau$ grows logarithmically with system size ($\tau = O(\log L)$), reflecting the time required for magic to propagate.
   • Long-Range Architectures: For circuits with long-range interactions (decay exponent $\alpha < 2$), magic can "jump" across the system, allowing the ensemble to reach the asymptotic design value in constant depth ($\tau = O(1)$).

3. Robustness to Injection Protocol:
   The paper demonstrates that the specific timing of magic injection (initial state preparation vs. measurement basis rotation) does not alter the fundamental behavior. Injecting magic via rotated measurements yields the same convergence dynamics as injecting it into the initial state, provided the local distribution of non-Clifford gates is equivalent.

4. Limits of Stabilizer States:
   The study confirms that without magic ($M_k=0$), the projected ensemble of Clifford-scrambled states is restricted to a 1-design, regardless of the depth of the Clifford circuit. A single T-gate (minimal magic) introduces non-trivial $k$-dependence but is insufficient to drive the ensemble to a genuine high-order design without sufficient system size and scrambling.

Significance
The paper provides a systematic route to generating quantum state designs in early fault-tolerant devices. It demonstrates that a small, controlled amount of magic is sufficient to generate highly random states, bridging the gap between the limitations of pure Clifford circuits and the requirements for quantum advantage. By characterizing the relationship between stabilizer entropy and design order, the work offers a quantitative framework for assessing the "randomness" of projected ensembles in near-term architectures. The identification of "magic teleportation" further elucidates how measurement-induced randomness can be harnessed to overcome locality constraints in shallow circuits, a critical insight for analog quantum simulators and error-corrected quantum processors.