Theoretical physicist (postdoc) at ULB Brussels. Quantum simulation, tensor networks & fractional Chern insulators Latest work: https://xitter.cf/t.co/WQZuSHglZ4
- Tweets146
- Following212
- Followers360
- Likes436
ALT (a) Central charge c extracted from the bipartite entanglement entropy using a scaling collapse for four legs and at most four bosons per site (Nmax=4). We find extended phases with one (c=1) and two (c=2) chiral bosonic edge modes. (b) These phases coincide with those found for a hard-core three-body repulsion (Nmax = 2). In both panels the gray region indicates very weak interaction strengths U/tx < 0.5 where for Nmax = 4 the central charge could not be deter- mined with the necessary accuracy.
ALT In panel a), the typical training procedure is displayed: we sample snapshots of the system, evaluate the energy, update the neural network parameters and repeat... Panel b) shows the training progress of the recurrent neural network (RNN) for an exemplary set of parameters. Panel c) shows the phase diagram for a bosonic Hofstadter model with long-range van der Waals interactions.
ALT BCC won a first price for mixed choir and prices for the "best performance of the compulsory work", the "best performance of a moronic folk song", the "beat developed and innovative program" and the highest rated choir (91% !🥳)
ALT Stability regime of the bosonic Pfaffian state in short ladders, as extracted through the particle entanglement spectrum of 4-boson ground states obtained through exact diagonalization. The yellow region indicates a finite entanglement gap with both particle cuts N_A=1 and N_A=2, with a counting consistent with the 4 particle Pfaffian (respectively 3 and 6 states below the entanglement gap). The ladder lengths (respectively Lx = 12, 10, 9, 8) are adjusted to the number of legs to have a sufficient number of ladder sites while keeping the computational complexity manageable.
ALT (a) Possible experimental realization of the proposed measurement protocol using a synthetic bilayer Hofstadter-Hubbard system with internal states (1). (2) On one side of the system, the edges are coupled to obtain an effective single layer system. (3) Measurements of off-diagonal elements of the one-particle reduced density matrix are possible using a local pulse protocol involving both layers. (b) Correlations extracted from 10.000 numerical snapshots without any post-processing and with experimentally relevant errorbars from shot-noise. The algebraic decay of the two-site correlations of the composite bosons (red squares) can be clearly distinguished from the exponential decay of the original bosonic correlations (blue circles). (c) Tensor-network diagram of the extended perfect sampling scheme which allows for numerical snapshots of composite boson correlations equivalent to the ones obtained in the proposed experiment.
ALT In panel a), the typical training procedure is displayed: we sample snapshots of the system, evaluate the energy, update the neural network parameters and repeat... Panel b) shows the training progress of the recurrent neural network (RNN) for an exemplary set of parameters. Panel c) shows the phase diagram for a bosonic Hofstadter model with long-range van der Waals interactions.
