The interest in possible applications of machine learning in physics has been growing exponentially for a while now and there seems to be a sea of literature. Couple of months ago we decided to review the literature and have weekly seminar about the papers we found interesting. There is a dedicated blog post for each of these on our group website.
Recently I have contributed with the discussion of the paper by E. M. Stoudenmire and D. J. Schwab: Supervised Learning with Quantum-Inspired Tensor Networks (arXiv:1605.05775 (2017)). In this paper the authors propose to train a tensor network with DMRG-like sweeps. You can read my full post here.
In our latest work, now on arXiv, we show how to use a convolutional neural network to extract physical parameters (even the quantum ones!) from experimental currents.
In my PhD I was generally concerned with monitoring and parameter estimation of quantum systems. These elements are crucial for efficiently functioning quantum devices, and, in difference from on-chip quantum operations, there is still a long way to go in terms of getting efficient readout at reasonable times. The ability to extract the maximum amount of information from an experimental record is therefore essential.
In practice, the experimental noise is sometimes so stubborn and viciously correlated that it may be really hard if not impossible to construct a quantum model that describes it. In our work we show that even for the cases where traditional parameter estimation methods do not work the convolutional network is a great solution to find the parameters governing the dynamics of the system.
I am happy to announce, that our paper “Conditioned spin and charge dynamics of a single-electron quantum dot” (I discussed it here) have been chosen as Editors’ Suggestion in Physical Review A.
Recently I finished my latest work that has been done in collaboration with my wonderful supervisor and Oxford experimental team and I would like to use this post to advertise it a bit in general terms. You can read it in full at arXiv.
The past quantum state method relies on a simple assumption: since in practical experimental situations you would like to monitor your system continuously and collect as much data as possible it makes sense to condition your probability not only what happened to your system BEFORE the time t (that is any given time for which you would like to make you probability prediction), but also AFTER the time t. In other words you use both the PAST and the FUTURE (from the point of the time, t, you are interested in) to make a probability prediction. This might sound a little bit sci-fi but as in general in quantum reality it is nothing too fancy, you basically just need to modify the Born rule a bit. The method was first proposed here and we used this kind of reasoning to argue stuff about correlation functions and improve fidelity of the teleportation protocol.
Here we took on the challenge to improve the experimental readout of the single electron quantum dot as well as modify existing techniques for parameter estimation. As it turns out, for typical experimental parameters, we are able to remove most of the noise and we are able to find time of each tunnelling event with super high precision. In addition to that we modified the Baum-Welch parameter estimation method and combined it with good old Bayesian to estimate both coherent and incoherent parameters under the same footing. So if you like quantum dots or you are just interested in quantum measurement theory in general, please have a look!