Machine Learning for Quantum Metrology

Seminar author:Martí Perarnau LLobet

Event date and time:05/20/2014 02:30:pm

Event location:IFAE seminar room

Event contact:

Quantum physics is an essentially probabilistic theory, and thus the precision of any measurement is limited. Given a set of independent measurements on N probes (for example, photons or atoms), this precision is bounded by the standard quantum limit, which scales as $1/\sqrt{N}$. Quantum effects, particularly entanglement, can enhance the precision to $1/N$, which is known as the Heisenberg limit. Such an enhancement also requires either (i) a global measurement on the probes or (ii) an adaptive measurement scheme. Whereas the latter is experimentally friendlier, devising optimal adaptive protocols is far from straightforward, and one typically relies on either clever guessing or brute-force numerical optimization. Recently a new approach has been suggested: machine learning, which replaces guesswork by a logical, fully-automatic, programmable routine. In this talk I will present two important machine learning algorithms, differential evolution and particle swarm optimization, and show their applicability for quantum metrology. This talk is based on [1-3].

[1] A. Hentschel and B. C. Sanders, Phys. Rev. Lett. 104, 063603 (2010).

[2] A. Hentschel and B. C. Sanders, Phys. Rev. Lett. 107, 233601 (2011).

[3] A. Lovett NB, Crosnier C, Perarnau-Llobet M, Sanders BC, Phys. Rev. Let. 110, 220501 (2013).