Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
107 tokens/sec
Gemini 2.5 Pro Premium
58 tokens/sec
GPT-5 Medium
29 tokens/sec
GPT-5 High Premium
25 tokens/sec
GPT-4o
101 tokens/sec
DeepSeek R1 via Azure Premium
84 tokens/sec
GPT OSS 120B via Groq Premium
478 tokens/sec
Kimi K2 via Groq Premium
213 tokens/sec
2000 character limit reached

Operational interpretation of the Stabilizer Entropy (2507.22883v1)

Published 30 Jul 2025 in quant-ph

Abstract: Magic-state resource theory is a fundamental framework with far-reaching applications in quantum error correction and the classical simulation of quantum systems. Recent advances have significantly deepened our understanding of magic as a resource across diverse domains, including many-body physics, nuclear and particle physics, and quantum chemistry. Central to this progress is the stabilizer R\'enyi entropy, a computable and experimentally accessible magic monotone. Despite its widespread adoption, a rigorous operational interpretation of the stabilizer entropy has remained an open problem. In this work, we provide such an interpretation in the context of quantum property testing. By showing that the stabilizer entropy is the most robust measurable magic monotone, we demonstrate that the Clifford orbit of a quantum state becomes exponentially indistinguishable from Haar-random states, at a rate governed by the stabilizer entropy $M(\psi)$ and the number of available copies. This implies that the Clifford orbit forms an approximate state $k$-design, with an approximation error $\Theta(\exp(-M(\psi))$. Conversely, we establish that the optimal probability of distinguishing a given quantum state from the set of stabilizer states is also governed by its stabilizer entropy. These results reveal that the stabilizer entropy quantitatively characterizes the transition from stabilizer states to universal quantum states, thereby offering a comprehensive operational perspective of the stabilizer entropy as a quantum resource.

Summary

We haven't generated a summary for this paper yet.

Dice Question Streamline Icon: https://streamlinehq.com

Follow-up Questions

We haven't generated follow-up questions for this paper yet.

X Twitter Logo Streamline Icon: https://streamlinehq.com

Don't miss out on important new AI/ML research

See which papers are being discussed right now on X, Reddit, and more:

“Emergent Mind helps me see which AI papers have caught fire online.”

Philip

Philip

Creator, AI Explained on YouTube