Day 1 – Monday 4th May 12pm – 2pm (UK Time)
12.00 – 12.10 Introduction
12.10 – 12.40 R. MacKay – (University of Warwick, Coventry)
12.40 – 13.10 T. Komatsuzaki – (Hokkaido University, Japan)
13.10 – 13.40 S. Keshavamurthy – (Indian Inst of Technology Kanpur, India)
Day 2 – Tuesday 5th May 12pm – 2pm (UK Time)
12.00 – 12.10 Introduction
12.10 – 12.40 A. Mancho – (Instituto de Ciencias Matemáticas, CSIC, Madrid)
12.40 – 13.10 V. J. Garcia Garrido – (Universidad de Alcala, Spain)
13.10 – 13.40 Ch. Skokos – (University of Cape Town, South Africa)
Day 3 – Wednesday 6th May 12pm – 2pm (UK Time)
12.00 – 12.10 Introduction
12.10 – 12.40 S. Farantos – (University of Crete, Greece)
12.40 – 13.10 P. Patsis – (RCAAM of Academy of Athens, Greece)
13.10 – 13.40 P. Cincotta – (Universidad Nacional de La Plata, Argentina)
Symposium – Title: Crowd-sourcing Machine Learning in NMR
Following on from the success of the Kaggle competition that Champs PDRA Lars Anderson Bratholm ran last year he is now organising a Symposium to be held on Thursday 5th March 2020.
Title, abstract and schedule below (subject to change):
Crowd-sourcing Machine Learning in NMR
The rise of machine learning (ML) has led to an explosion in potential strategies which may be used to learn from data in order to make scientific predictions. For physical scientists who wish to apply ML strategies to a particular domain, the vast number of strategies available has made it difficult to make an a priori assessment of what strategy to adopt. This is further complicated when similar domains have not been previously explored in the literature.
Recently, we worked with Kaggle to design a competition which encouraged data scientists around the world to develop ML models for predicting pairwise nuclear magnetic resonance (NMR) properties for synthetically relevant chemical compounds. Over 3 months, we received 47,800 ML model submissions from 2700 teams in 84 countries, with the top models outperforming our own previously published methods. The success this strategy has cultivated highlights the potential of crowd-sourced ML approaches across a range of scientific domains.
This symposium will introduce the background and main findings of the competition, including the context of computational NMR, the Kaggle platform and presentations from the top performing teams of the competition.
10:00 – 11:15: Registration
11:15 – 11:50: Craig Butts and Will Gerrard
11:50 – 12:25: Addison Howard and Walter Reade
12:25 – 13:15: Lunch
13:15 – 13:50: Lars Andersen Bratholm
13:50 – 14:25: Brandon Anderson
14:25 – 14:40: Break
14:40 – 15:15: Luka Stojanovic and Milos Popovic
15:15 – 15:50: Youhan Lee and Sunghwan Choi
15:50 – 16:05: Break
16:05 – 16:40: Andres Torrubia
16:40 – 17:15: Devin Wilmott
17:15 – 18:15: Drinks reception An Eventbrite registration, link below, has been set up. Numbers are limited (due to lecture theatre capacity) and will be allocated on a first come, first served basis. If you wish to attend please can you complete the form so we can plan for numbers and catering on the day.
On December 6-9, 2019, CHAMPS PDRAs Makrina Agaoglou, Broncio Aguilar Sanjuan, Rafael Garcia-Meseguer, Francisco Gonzalez-Montoya, Matthaios Katsanikas, Vladimir Krajnak, Shibabrat Naik, PI Stephen Wiggins, and collaborator Victor Jose Garcia-Garrido of the Universidad de Alcala in Spain held a book sprint.
A book sprint is a method of creating a book collaboratively in a short period of time. The book sprint was held at Engineers House in Bristol which afforded an atmosphere promoting intensive collaboration with minimal distractions. The book is an account of our results, approach, and future directions on the phase space approach to chemical reaction dynamics. The book was produced using Jupyter Book and is available at www.chemicalreactions.io. The result is the book entitled “Chemical Reactions: A Journey into Phase Space”. It is our intention that the book will appeal both to mathematicians and to chemists and will serve as an entry into the field. Indeed, the GitHub framework provides a mechanism for scientific collaboration and we invite the interested scientific community to contribute to the further development of the book. Instructions for this are on the GitHub site for the book (https://github.com/champsproject/chem_react_dyn).
On Wednesday the 18th of September a joint CHAMPS and Leeds Physical Chemistry seminar “Quantum Trajectories in Phase Space” took place in Leeds . The talk was given by Prof Craig Martens from University of California, Irvine. Although it is a physical/computational chemistry seminar it was of general interest, as it was focused on unusual formulation of quantum mechanics. We almost always think about quantum mechanics in terms of wave function and waves of probability, but there are other ways of looking at it. But Prof Martens presented one an approach, in which quantum mechanics is formulated very similar to classical mechanics, with the difference that neighbouring classical trajectories are not independent. They are entangled and push each other in a very specific way, as if there were many parallel worlds interacting with each other. The talk was attended not only by physicist, mathematicians and computational chemists, but also by a number of experimental organic and inorganic chemists.
Although it is a physical/computational chemistry seminar it can be of general interest, as it will be about an unusual formulation of quantum mechanics. We almost always think about quantum mechanics in terms of wave function and waves of probability, but there are other ways of looking at it. Craig Martens will present one of those approaches, in which quantum mechanics is just like classical, with the difference that neighbouring classical trajectories are not independent. They are entangled and push each other in a very specific way, a bit like there were many parallel worlds interacting with each other.
The CHAMPS Project was pleased to host Professor Joel Bowman, the Samuel Candler Dobbs Professor of Theoretical Chemistry at Emory University in Atlanta, Georgia, for a visit during September 3-5, 2019. Professor Bowman presented a seminar entitled “A Machine Learning Approach for Prediction of Rate Constants”, which described a new application of machine learning in chemistry. Actually, the talk had two independent halves, with the first half of the talk devoted to a discussion of the current “state-of-the art” of the roaming mechanism for chemical reactions, which included some discussion of the manifestation of quantum effects in roaming .The CHAMPS PDRAs (as well as the PI) particularly enjoyed a two hour, informal, round table discussion, covering many aspects of contemporary reaction dynamics, as well as a bit of much appreciated career advice.
Figuring premier @ the Wickham Theatre! Posted on September 30, 2018 by David R Glowacki
The last couple weeks have seen some interesting outreach developments! With support from Arts Council England, the EPSRC-funded CHAMPS programme, the Leverhulme Trust, and the Royal Society, David Glowacki and team worked on a project called “Figuring” at the Wickham theatre in the University of Bristol’s Department of Drama, with a talented team drawn across artistic, scientific, and technological practices. On 21 Sept, they worked to premier Figuring to an audience of artists, producers, and technologists. This follows on from a previous prototype showing of Figuring at the Knowle West Media Centre, as part of their Commons Sense programme.
The aim of Figuring is to investigate what can be created when moving, sensing bodies are embedded in simulated virtual worlds, and what arises when somatic and movement based practices are combined with Narupa, a state-of-the-art multi-person VR framework for molecular dynamics simulation, which has been developed within the Intangible Realities Laboratory over the last several years.
Figuring takes its name from its intention to explore ‘string figures’, in both the real world and also in the virtual world. String figures are created through simple movements of folding, looping, twisting, and knotting strings between the hands, fingers and thumbs of one or more people.
During our time at the Wickham theatre, members of the CHAMPS project worked with a group of dancers to facilitate experiments with both physical and virtual strings. For the ‘raw material’ of the virtual strings, the team relied on real-time molecular dynamics simulations of proteins: the molecular strings from which life is woven. Narupa enables audiences to reach out and touch simulated proteins: folding, looping, twisting, and knotting them.
Despite their virtuality, Figuring audiences reported ‘felt’ sensations whilst manipulating virtual simulations of molecular strings. Moving forward, we hope to better understand the origins of such sensations, how they map onto their physical and tangible analogues, and how different sensory and somatic practices might enable us to understand perception across real, virtual and imagined environments.
Figuring represents a collaboration between a diverse team with broad interests, led by David Glowacki and somatic/movement artist Lisa May Thomas. Alex Jones, Dr. Tom Mitchell, and Prof. Joseph Hyde helped to devise algorithms for generating sound from the virtual string dynamics. Computer Scientist Mike O’Connor and Mark Wannacott played a key role in developing the VR interaction capabilities and aesthetic, and Helen Deeks provided key advice on human-computer interaction strategies. Somatic and movement practitioners included Laila Diallo, Ben McEwan, Bryn Thomas, Ania Varez, Will Dickie, Fernanda Munoz-Newsome and Anne-Gaëlle Thiriot. Production is by Emma Hughes, dramaturgy by Tanuja Amarasuriya, and set design by Phillipa Thomas. Photos are by Paul Blakemore and Silvia Cardarelli-Gronau, and film by Adam Laity.
Summer Undergraduate Research Opportunities in the CHAMPS Project
The breadth and scale of research in CHAMPS provides opportunities for researchers from a variety of backgrounds and levels, even for undergraduates. This summer Wenyang Lyu , an undergraduate in the School of Mathematics at the University of Bristol, personified these characteristics of CHAMPS. Supported by a bursary from the London Mathematical Society and the School of Mathematics, Wenyang completed a research project that will be useful for the CHAMPS project, as well as researchers in related areas worldwide.
Periodic orbits are universally acknowledged as a fundamental building block for phase space structure in dynamical systems. For general nonlinear dynamical systems discovering the existence and nature of periodic orbits can only be carried out using computational methods. For this reason it is important to have numerical methods for fining periodic orbits that lead to reproducible results. Wenyang implemented two know numerical methods for finding periodic orbits in Python, and further developed a new numerical method. He has tested these methods on benchmark two degree-of-freedom Hamiltonian systems.
Phase space structures such as dividing surfaces, normally hyperbolic invariant manifolds, and their stable and unstable manifolds in molecular Hamiltonian have been an integral part of computing quantitative results such as the cumulative reaction probability and rate constants in chemical reactions. Thus, methods that can reveal the geometry of these invariant manifolds in high dimensional phase space (4 or more dimensions) need to be benchmarked by comparing with known results. In these articles, we assessed the capability of one such method called Lagrangian descriptor (LD) for revealing the aforementioned high dimensional phase space structures associated with an index-1 saddle in Hamiltonian systems. The LD based approach is applied to two and three degree-of-freedom quadratic Hamiltonian systems where the high dimensional phase space structures are known, that is as closed-form analytical expressions. This leads to a direct comparison of features in the LD contour maps and the phase space structures’ intersection with an isoenergetic two-dimensional surface, and hence provides a verification of the method. Next, the method of LD is applied to classical two and three degrees of freedom Hamiltonians that model features of dissociation reactions. The result of the LD based approach is compared with an established numerical method for computing unstable periodic orbit and tube manifolds of a two degrees of freedom system, and the results are in good agreement. We have also discussed the results in the context of three degrees of freedom extension of the same model Hamiltonian. References:
Shibabrat Naik, Víctor J. García-Garrido, Stephen Wiggins, Finding NHIM: Identifying high dimensional phase space structures in reaction dynamics using Lagrangian descriptors, Communications in Nonlinear Science and Numerical Simulation, 2019, 79, 104907 https://doi.org/10.1016/j.cnsns.2019.104907
Shibabrat Naik and Stephen Wiggins, Finding normally hyperbolic invariant manifolds in two and three degrees of freedom with Hénon-Heiles-type potential, Phys. Rev. E, 2019, 100 (2), 022204 https://doi.org/10.1103/PhysRevE.100.022204