Given an unknown quantum state rho, and a known list of two-outcome measurements E_1,…,E_M, \u00abshadow tomography\u00bb is the task of estimating the probability that each E_i accepts rho, by carefully measuring only a few copies of rho. In 2018, I gave the first nontrivial protocol for this task. In 2019, Guy Rothblum and I exploited a new connection between gentle measurement of quantum states and the field of differential privacy, to give a protocol that requires fewer copies of rho in some cases, and has the additional advantage of being online (that is, the measurements are processed one at a time). Huge challenges remain in making shadow tomography practical with near-term devices; extremely recently Huang, Kueng, and Preskill took some promising steps in that direction. I’ll survey these developments and the challenges that remain. <\/p>\n
Papers: <\/p>\n
https:\/\/www.scottaaronson.com\/ https:\/\/www.scottaaronson.com\/