Reproduce “Agnostic Process Tomography” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chirag Wadhwa, Laura Lewis, Elham Kashefi, Mina Doosti
Researchers at the University of Edinburgh define Agnostic Process Tomography (APT), an extension of agnostic quantum state learning, to characterize quantum processes with simple structures even when the true process is noisy or deviates from ideal assumptions. They developed algorithms for various concept classes, achieving proper 1-agnostic learning with O(1) sample and time complexities for Pauli strings and channels, and established methods for approximating complex unitaries and channels with simpler, structured counterparts.
Key insights
- Agnostic learning can be effectively extended from quantum states to quantum processes, providing robustness against noise and structural deviations in practical quantum settings.
- For certain simple concept classes like Pauli strings and Pauli channels, highly efficient agnostic learning algorithms (O(1) complexity in system size) can be developed.
- The distinction between sample efficiency (query complexity) and computational efficiency (time complexity) is critical in agnostic tomography, with some sample-efficient methods remaining computationally intensive.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.