Séminaire de Physique ThéoriqueTensor Networks and continuous Entanglement Renormalisation
Jos Van Doorsselaere (Université de Gand, Belgique)
Thursday 11 May 2017 14:00 - Tours - Salle 1180 (Bât E2)
We will give a general introduction to tensor networks end specifically the MERA. Tensor networks are an efficient alternative to Monte-Carlo simulation to compute ground states of strongly correlated quantum (spin) systems. A particular implementation is the MERA, which gives the ground state of a 1D system, but organises the links in the network according to scale in an extra ‘dimension’ and thus has a build in renormalisation group structure. Next we will illustrate the nice interplay between the tensor networks and field theories. The renormalisation properties of a MERA are hard to extend to continuous systems. One approach is looking at similarities with holographic theories, relying on the Ryu-Takayanagi property that geodesic lengths are a measure for entanglement entropy. Another more straightforward approach is to look at interacting fermionic duals of free bosonic systems.