English

Improved quantum data analysis

Quantum Physics 2024-08-07 v4 Computational Complexity

Abstract

We provide more sample-efficient versions of some basic routines in quantum data analysis, along with simpler proofs. Particularly, we give a quantum "Threshold Search" algorithm that requires only O((log2m)/ϵ2)O((\log^2 m)/\epsilon^2) samples of a dd-dimensional state ρ\rho. That is, given observables 0A1,A2,...,Am10 \le A_1, A_2, ..., A_m \le 1 such that tr(ρAi)1/2\mathrm{tr}(\rho A_i) \ge 1/2 for at least one ii, the algorithm finds jj with tr(ρAj)1/2ϵ\mathrm{tr}(\rho A_j) \ge 1/2-\epsilon. As a consequence, we obtain a Shadow Tomography algorithm requiring only O~((log2m)(logd)/ϵ4)\tilde{O}((\log^2 m)(\log d)/\epsilon^4) samples, which simultaneously achieves the best known dependence on each parameter mm, dd, ϵ\epsilon. This yields the same sample complexity for quantum Hypothesis Selection among mm states; we also give an alternative Hypothesis Selection method using O~((log3m)/ϵ2)\tilde{O}((\log^3 m)/\epsilon^2) samples.

Keywords

Cite

@article{arxiv.2011.10908,
  title  = {Improved quantum data analysis},
  author = {Costin Bădescu and Ryan O'Donnell},
  journal= {arXiv preprint arXiv:2011.10908},
  year   = {2024}
}