<p>Entanglement entropy (EE) is a quantum information theoretic measure that quantifies the correlations between a region and its surroundings. We study this quantity in the following two setups : </p> <ul> <li>We look at the dynamics of a free minimally coupled, massless scalar field in a deSitter expansion, where the expansion stops after some time (i.e. we quench the expansion) and transitions to flat spacetime. We study the evolution of entanglement entropy (EE) and the Rényi entropy of a spatial region during the expansion and, more interestingly, after the expansion stops, calculating its time evolution numerically. The EE increases during the expansion but the growth is much more rapid after the expansion ends, finally saturating at late times, with saturation values obeying a volume law. The final state of the subregion is a partially thermalized state, reminiscent of a Gibbs ensemble. We comment on application of our results to the question of when and how cosmological perturbations decohere.</li> <li>We study the EE in a theory that is holographically dual to a BTZ black hole geometry in the presence of a scalar field, using the Ryu-Takayangi (RT) formula. Gaberdiel and Gopakumar had conjectured that the theory of N free fermions in 1+1 dimensions, for large N, is dual to a higher spin gravity theory with two scalar fields in 2+1 dimensions. So, we choose our boundary theory to be the theory of N free Dirac fermions with a uniformly winding mass, m e^iqx, in two spacetime dimensions (which describes for instance a superconducting current in an N-channel wire). However, to O(m²), thermodynamic quantities can be computed using Einstein gravity. We aim to check if the same holds true for entanglement entropy (EE). Doing calculations on both sides of the duality, we find that general relativity does indeed correctly account for EE of single intervals to O(m²).</li> </ul>

Mes recherches portent essentiellement sur l''elaboration de m'ethodes de calcul formel pour l''etude constructive des 'equations diff'erentielles lin'eaires, plus particuli'erement autour de la th'eorie de Galois diff'erentielle. Celles-ci vont du d'eveloppement de la th'eorie sous-jacente aux algorithmes, en incluant leur implantation en Maple. Ces travaux ont en commun une approche exp'erimentale des math'ematiques o'u l'on met l'accent sur l'examen d'exemples les plus pertinents possibles. L''etude d'etaill'ee de cas provenant de la m'ecanique rationnelle ou de la physique th'eorique nourrit en retour le d'eveloppement de th'eories math'ematiques idoines. Mes travaux s'articulent suivant trois grands th'emes interd'ependants : la th'eorie de Galois diff'erentielle effective, ses applications 'a l'int'egrabilit'e de syst'emes hamiltoniens et des applications en physique th'eorique.