In April 2015, CECAM initiated a series of lectures to highlight interesting developments in different areas of computational science. Each year, two keynote speakers are invited to present their recent work. The lectures are held in the EPFL campus with the goal to enhance the opportunities for scientific exchange offered by CECAM's workshop and conference program by sharing some of its most exciting topics with the broader EPFL community.
Almost famous, a woman behind the codes
Many of the breakthroughs of the early days of simulation would not have been possible without skilled programmers who translated new scientific ideas into efficient codes that would run without errors on the supercomputers of the 1950s and 1960s. On November 15 2017, CECAM and NCCR MARVEL invited a truly outstanding representative of the first generation of coders, Mary Ann Mansigh Karlsen, to share her recollections of those pioneering efforts. Mary Ann's long collaboration with Berni Alder at Lawrence Livermore National Laboratory was instrumental for establishing molecular dynamics as a key tool to study condensed matter systems. The event was introduced by a presentation on the early days of computer simulations in condensed matter physics by former CECAM director Prof. Michel Mareschal.
Below are the videos of the two parts of the event.
Part 1: It's only numerics: an introduction to the early days of computer simulation by Michel Mareschal
Part 2: A conversation between Mary Ann Mansigh Karlsen and Daan Frenkel (University of Cambridge, UK). Insights on her fascinating story and her experiences as a coder at Livermore over a period of almost thirty years, providing a unique point of view on the scientific and technological development of computer simulation from the early days of molecular dynamics in the late 1950, to the mid 1980s.
CECAM Lecture: Prof. Erik Lindahl Modeling Experiments in Simulations and Simulations in Experiments: Combining Biomolecular Simulations with Low-Resolution Measurements, Wednesday 23 November 2016 at 16:30, Auditorium CM2.
Abstract. Modeling and simulations of complex biological macromolecules has made tremendous progress since the 1960s as the result of a wonderful marriage of theoretical and computational advances. However, while molecular dynamics simulations have been wonderful at exploring and explaining the atomic detail of processes, with a few exceptions it is not until the last few years these methods have become predictive enough to compete with experimental techniques. I will discuss some of these challenges, such as whether force fields are accurate enough and if we should reconsider some of the decisions made in the 60s in the light of computational advances. Can we find better ways to combine molecular dynamics with experimental measurements to gain insight? I will discuss how simulations have been used by a number of teams to mimic and reproduce experimental measurements in innovative ways (as well as some failures, in particular our own ones). Experiments are also increasingly designed to specifically probe simulation results, which opens new exciting opportunities. How is low-resolution or even time-dependent experimental data best integrated into simulations, why is it useful, and where should we head in the near future?
CECAM Lecture: Prof. Christof Schütte Computational Molecular Design: Mathematical Theory, High Performance Computing, In Vivo Experiments , Monday 25 May 2016 at 17:00, Auditorium CM1.
Abstract. Molecular dynamics and related computational methods enable the description of biological systems with all-atom detail. However, these approaches are limited regarding simulation times and system sizes. A systematic way to bridge the micro-macro scale range between molecular dynamics and experiments is to apply coarse-graining (CG) techniques. We will discuss Markov State Modelling, a CG technique that has attracted a lot of attention in physical chemistry, biophysics, and computational biology in recent years. First, the key ideas of the mathematical theory and its algorithmic realization will be explained, next we will discuss the question of how to apply it to understanding ligand-receptor binding, and last we will ask whether this may help in designing ligands with prescribed function. All of this will be illustrated by telling the story of the design process of a pain relief drug without concealing the potential pitfalls and obstacles.
CECAM Lecture: Prof. Eberhard Gross Analysis and control of electron dynamics: An ab-initio perspective on the femto-second time scale, Monday 28 September 2015 at 17:00, Auditorium CM1.
Abstract. This lecture is about the motion of electrons, how it can be monitored, analyzed and, ultimately, controlled with external fields on the femto-second time scale. The investigations are performed with ab-initio simulations, using time-dependent density functional theory as theoretical tool. We shall visualize the laser-induced formation and breaking of chemical bonds in real time, and we shall adress questions like: How much time needs an electron to complete a transition from one state to another? Another main topic will be quantum transport. Time-dependent features of the electronic current through nano-scale junctions will be studied for electron pumps and molecular optical switches. A combination of quantum optimal control theory with time-dependent density functional theory will be presented as a method to compute laser pulses that are optimised to achieve a given goal. As an example, we shall calculate the laser pulse needed to switch the chirality of currents in quantum rings. Finally we will study the ultrafast laser induced demagnetisation of ferromagnetic solids.
CECAM lecture: Prof. Ali Alavi Do we really need Quantum Computers to simulate Quantum Chemistry? Monday 22 Avril 2015 at 17:00
Abstract. The accurate calculation of the ground state of many-electron systems has been the central goal of quantum chemistry for the last 80 years. Exact methods, such as full CI, can only be applied to systems of a few electrons and it has long been assumed that larger fermionic systems will only be simulated exactly on powerful "quantum computers". Here we argue that this is not the case for a large class of realistic electronic systems, with up to ~50 electrons. The ground states of these fermion systems can be calculated using a very simple stochastic algorithm, based on a population dynamics of a set of annihilating walkers of positive and negative sign in the space of the Slater determinants of the system. We show that this algorithm can be used to solve difficult fermion systems to unprecedented accuracy, as exemplified by a recent application to the ionisation potential of the first row (3d) transition metal atoms. Furthermore, we show that a replica trick allows the unbiased calculation of two-particle correlation functions, as well as excited states.