Différences
Ci-dessous, les différences entre deux révisions de la page.
| Les deux révisions précédentes Révision précédente Prochaine révision | Révision précédente | ||
|
animation:seminaires:2011:accueil [2017/03/10 11:08] cicaluga [Séminaires 2011] |
animation:seminaires:2011:accueil [2017/03/10 11:09] (Version actuelle) cicaluga [Séminaires 2011] |
||
|---|---|---|---|
| Ligne 48: | Ligne 48: | ||
| * Jean-Philippe MAGUE, Humanités Numériques, Ecole Normale Supérieure de Lyon | * Jean-Philippe MAGUE, Humanités Numériques, Ecole Normale Supérieure de Lyon | ||
| - | (24 participants) | + | **(24 participants)** |
| Abstract: Les sciences humaines et sociales sont restées jusqu'à récemment, dans leur grande partie, peu, voire pas du tout, équipées en outils et instruments. La généralisation de l'accès à des données numérisées et la quantité croissante de ces données changent cet état de fait : le format numérique rend possible l'utilisation d'outils pour conduire l'analyse, et ces derniers sont même une condition nécessaire au passage à l'échelle de ces analyses. Ce mouvement de fond d'outillage et d'instrumentation, qui change le rapport que le chercheur entretient avec son objet d'étude, est transversal à l'ensemble des SHS et est couramment désigné par le terme d'Humanités Numériques. Nous proposerons un parcours à travers cette "transdiscipline" émergente. | Abstract: Les sciences humaines et sociales sont restées jusqu'à récemment, dans leur grande partie, peu, voire pas du tout, équipées en outils et instruments. La généralisation de l'accès à des données numérisées et la quantité croissante de ces données changent cet état de fait : le format numérique rend possible l'utilisation d'outils pour conduire l'analyse, et ces derniers sont même une condition nécessaire au passage à l'échelle de ces analyses. Ce mouvement de fond d'outillage et d'instrumentation, qui change le rapport que le chercheur entretient avec son objet d'étude, est transversal à l'ensemble des SHS et est couramment désigné par le terme d'Humanités Numériques. Nous proposerons un parcours à travers cette "transdiscipline" émergente. | ||
| Ligne 58: | Ligne 58: | ||
| **March 15, 2011 - 14:00** | **March 15, 2011 - 14:00** | ||
| * Tristan ALBARET, Laboratoire de Physique de la Matière Condensée et Nanostructures, Université Claude Bernard Lyon 1 | * Tristan ALBARET, Laboratoire de Physique de la Matière Condensée et Nanostructures, Université Claude Bernard Lyon 1 | ||
| + | |||
| + | **(26 participants)** | ||
| Abstract: For many material science systems, an accurate enough atomistic description cannot be obtained either by full quantum approaches based on electronic structure calculations nor by classical schemes using empirical potentials. These difficulties arise from the limited system size imposed by the high computational cost of the quantum schemes and by the lack of accuracy of the empirical potentials when they are employed in conditions that depart strongly enough from those for which they have been designed. However, in a large class of problems, the regions where the "quantum" accuracy is required is small enough and relatively well identified. In such a situation it may be worthwhile to describe concurrently the quantum and classical regions within a single hybrid scheme that realizes the coupling. This is the aim of the so-called QM/MM techniques which have been developed in various fields as for instance bio-chemistry or material science. In this presentation I will focus on material science problems for which we have developed our own technique called "Learn On The Fly". First, I will explain the technique and then discuss its application to the brittle fracture in Silicon. I will show that the results provide a consistent description of several experimental observations concerning crack propagation instabilities and the structures of the cleaved surfaces. Before concluding, and if time allows it, I will briefly talk about the main drawbacks of the LOTF technique and also about its possible extensions to other systems and problems like, for instance, the "acceleration" of ab-initio molecular dynamics. | Abstract: For many material science systems, an accurate enough atomistic description cannot be obtained either by full quantum approaches based on electronic structure calculations nor by classical schemes using empirical potentials. These difficulties arise from the limited system size imposed by the high computational cost of the quantum schemes and by the lack of accuracy of the empirical potentials when they are employed in conditions that depart strongly enough from those for which they have been designed. However, in a large class of problems, the regions where the "quantum" accuracy is required is small enough and relatively well identified. In such a situation it may be worthwhile to describe concurrently the quantum and classical regions within a single hybrid scheme that realizes the coupling. This is the aim of the so-called QM/MM techniques which have been developed in various fields as for instance bio-chemistry or material science. In this presentation I will focus on material science problems for which we have developed our own technique called "Learn On The Fly". First, I will explain the technique and then discuss its application to the brittle fracture in Silicon. I will show that the results provide a consistent description of several experimental observations concerning crack propagation instabilities and the structures of the cleaved surfaces. Before concluding, and if time allows it, I will briefly talk about the main drawbacks of the LOTF technique and also about its possible extensions to other systems and problems like, for instance, the "acceleration" of ab-initio molecular dynamics. | ||
| Ligne 67: | Ligne 69: | ||
| **February 15, 2011 - 15:00** \\ | **February 15, 2011 - 15:00** \\ | ||
| * Stefano RUFFO, Dipartimento di Energetica "S. Stecco", Università di Firenze, and Chair d'Excellence ENS-Lyon | * Stefano RUFFO, Dipartimento di Energetica "S. Stecco", Università di Firenze, and Chair d'Excellence ENS-Lyon | ||
| + | |||
| + | **(22 participants)** | ||
| Abstract: Systems with long-range interactions, like gravitational, charged and dipolar systems, can be made extensive, but are intrinsically non additive. The violation of this basic property of thermodynamics is the origin of ensemble inequivalence, which in turn implies that specific heat can be negative in the microcanonical ensemble, temperature jumps can appear at microcanonical first order phase transitions, ergodicity may be broken. Realizing that such features may be present for a wide class of systems has renewed the interest in long-range interactions. In this seminar, I will present a review of the recent advances on the statistical mechanics and out-of-equilibrium dynamics of models with long-range interactions. | Abstract: Systems with long-range interactions, like gravitational, charged and dipolar systems, can be made extensive, but are intrinsically non additive. The violation of this basic property of thermodynamics is the origin of ensemble inequivalence, which in turn implies that specific heat can be negative in the microcanonical ensemble, temperature jumps can appear at microcanonical first order phase transitions, ergodicity may be broken. Realizing that such features may be present for a wide class of systems has renewed the interest in long-range interactions. In this seminar, I will present a review of the recent advances on the statistical mechanics and out-of-equilibrium dynamics of models with long-range interactions. | ||
| Ligne 76: | Ligne 80: | ||
| **January 25, 2011 - 14:00** \\ | **January 25, 2011 - 14:00** \\ | ||
| * Stéphane LABROSSE, Laboratoire des Sciences de la Terre, ENS de Lyon , France | * Stéphane LABROSSE, Laboratoire des Sciences de la Terre, ENS de Lyon , France | ||
| + | |||
| + | **(26 participants)** | ||