arXiv subjects

Oscar Dahlsten

Publications and source records attributed to Oscar Dahlsten.

42 records · Page 3Linked to original sources

arXiv2016

Quantum generalisation of feedforward neural networks

We propose a quantum generalisation of a classical neural network. The classical neurons are firstly rendered reversible by adding ancillary bits. Then they are generalised to being quantum reversible, i.e.\ unitary. (The classical networks we generalise are called feedforward, and have step-function activation functions.) The quantum network can be trained efficiently using gradient descent on a cost function to perform quantum generalisations of classical tasks. We demonstrate numerically that it can: (i) compress quantum states onto a minimal number of qubits, creating a quantum autoencoder, and (ii) discover quantum communication protocols such as teleportation. Our general recipe is theoretical and implementation-independent. The quantum neuron module can naturally be implemented photonically.

quant-ph↗

arXiv2015

Photonic Maxwell's demon

We report an experimental realisation of Maxwell's demon in a photonic setup. We show that a measurement at the single-photon level followed by a feed-forward operation allows the extraction of work from intense thermal light into an electric circuit. The interpretation of the experiment stimulates the derivation of a new equality relating work extraction to information acquired by measurement. We derive a bound using this relation and show that it is in agreement with the experimental results. Our work puts forward photonic systems as a platform for experiments related to information in thermodynamics.

quant-ph↗

arXiv2013

Comment on "Quantum Szilard Engine"

In a recently published letter [S. W. Kim, T. Sagawa, S. DeLiberato, and M. Ueda, PRL 106, 070401 (2011)] the influence of particle statistics on extractable work in the Szilard engine was discussed. We point out that the expressions given there suggest no work extraction is possible in the low temperature limit if more than two particles are used and thus are not optimal. We argue that the optimal extractable work is in general higher and in particular non-decreasing in the number of particles.

quant-ph↗

arXiv2012

Measurement and Particle Statistics in the Szilard Engine

A Szilard Engine is a hypothetical device which is able to extract work from a single thermal reservoir by measuring the position of particles within the engine. We derive the amount of work that can be extracted from such a device in the low temperature limit. Interestingly, we show this work is determined by the information gain of the initial measurement rather than by the number and type of particles which constitute the working substance. Our work provides another clear connection between information gain and extractable work in thermodynamical processes.

quant-ph↗

arXiv2011

The thermodynamic meaning of negative entropy

Landauer's erasure principle exposes an intrinsic relation between thermodynamics and information theory: the erasure of information stored in a system, S, requires an amount of work proportional to the entropy of that system. This entropy, H(S|O), depends on the information that a given observer, O, has about S, and the work necessary to erase a system may therefore vary for different observers. Here, we consider a general setting where the information held by the observer may be quantum-mechanical, and show that an amount of work proportional to H(S|O) is still sufficient to erase S. Since the entropy H(S|O) can now become negative, erasing a system can result in a net gain of work (and a corresponding cooling of the environment).

quant-ph↗

arXiv2007

Exact entanglement probability distribution of bi-partite randomised stabilizer states

We study the entanglement properties of random pure stabilizer states in spin-1/2 particles. For two contiguous groups of spins of arbitrary size we obtain a compact and exact expression for the probability distribution for the entanglement values between the two sets. This allows for exact derivations of the average entanglement and the degree of concentration of measure around this average. We also give simple bounds on these quantities. We find that for large systems the average entanglement is near maximal and the measure is concentrated around it.

quant-ph↗