26.03.2026
17:45
HIT F31.1
26.03.2026
17:45
HIT F31.1
Dr. Yuxuan Zhang
Post-Doc at EPFL
Abstract
Random circuit sampling (RCS) has become a central framework for demonstrating quantum computational advantage, underlying recent experiments by Google, USTC, and others. I will begin with a brief introduction to RCS and the basic complexity-theoretic ideas supporting its hardness. A key practical ingredient in these experiments is cross-entropy benchmarking (XEB), which serves as a proxy for fidelity in the anti-concentrated (Porter–Thomas) regime. However, I will highlight two important limitations: XEB is difficult to scale and can be vulnerable to classical spoofing strategies. This motivates the study of peaked circuit sampling , where the output distribution has additional structure beyond the standard random ensemble. I will discuss how this setting offers a different route to quantum advantage, and what new challenges it raises for practical usage.