Good morning,
We have two talks in this week's seminar, by Anna Wegloop and Nobert Lütkenhaus. The talks will be in HIT on Hoenggerberg.
Tuesday, March 29, 16:30 - 18:00, HIT J 51.
Best regards, Stefan
___________________________________
Anna Wegloop
Parameter estimation using Minimum Probability Flow Learning
Abstract: It is often computationally expensive or even impossible to calculate the normalization factor of probability density functions. This hinders parameter estimation for statistical models. Minimum Probability Flow Learning and Score Matching are methods in which the normalization factor does not have to be calculated. Minimum Probability Flow Learning is a new method based on concepts from statistical mechanics. The method will be introduced and it it will be shown that Score Matching is a special case of this method.
___________________________________
Nobert Lütkenhaus
Testing Quantum Devices for Quantum Communication
Quantum Communication requires a number of different devices: quantum channels, quantum memories, quantum repeaters. As experimental effort is under way to realize these tools, we will necessarily start with imperfect devices. Channel loss is one example, decoherence in quantum memory another.
In my group we study simple test procedure to make sure that quantum devices can be of use for quantum communication. The criteria being that the performance of the devices cannot be explained by a process that measures out the input state in order to re-prepare the output from classical information (measure/reprepare strategy). This approach links the performance of devices as quantum devices to the ability to maintain entanglement. The technical key point will be methods to verify and quantify entanglement on quantum systems which consist of a combination of a discrete variable and a continuous variable system with just a handful of observed expectation values.
We concentrate on optical devices which operate in infinite dimensional Hilbert spaces and utilize simple test state, such as laser pulses, and quadrature measurements. We are able to show that a three different test states can be as powerful as an infinite set of coherent states with a Gaussian distribution of amplitudes to demonstrate quantum behavior.
In a quantitative approach, we give bounds on the amount of entanglement that can be distributed and/or stored with the devices.
___________________________________