Machine Learning meets Modelling and Simulation Methods

Within the new phase of CompBioMed we have dedicated a new Work Package to the emerging topic of machine learning and artificial intelligence with a focus on high performance data analytics required for these approaches. This meeting will be an opportunity for certain Core and Associate Partners working on this topic to gather and discuss current progress and the possibilities and needs to advance this burgeoning field.
The agenda incorporates talks from those in the field and those looking to progress into the field, followed by a day of software development sessions to advance the work and to learn from one another.

Last update: March 26th 18:00
The meeting was conducted through a virtual conference and all talks and presentations were recorded. Most of the slides and videos are now available here and on YouTube.

Day 1 – Monday 16th March 2020
12:00 – 13:30 Registration
13:30 – 13:35 Introduction
13:35 – 14:15 The Interplay of Machine Learning and Molecular Dynamics

Slides | Video

Keynote Talk: Ben Leimkuhler, University of Edinburgh
14:15 – 15:15 Drug discovery & personalised medicine: searching chemical space + free energy calculations Ola Engkvist, AstraZeneca

Slides | Video

Artificial Intelligence in Drug Design, Progress and Bottlenecks
Tristan Bereau, MPG/UvA

Slides | Video

Exploring chemical space with multiscale simulations
Austin Clyde (Argonne National Laboratory)

Slides | Video

Duelling GPUs: Scanning Chemical Space with Coupled Generative Models and Property Models
15:15 – 15:45 Break
15:45 – 17:25 Drug discovery & personalised medicine

Combining ML and MD for controlled evolution of macromolecular dynamics</strong

Organ level modelling: imaging + combination with simulation

Gianni De Fabritiis, UPF

(Confidential)

From functional to machine learning potentials in molecular simulations
Philippe Hupé, Institut Curie

Slides | Video

Software development and optimisation of bioinformatics pipelines to analyse high-throughput sequencing data in oncology
Andrew Potterton, UCL

Slide | Video

Using machine learning to understand the molecular properties that confer extended drug-target residence time
Shantenu Jha, Rutgers DeepDrive: Deep-Learning Driven Adaptive Molecular Simulations
Vicente Grau, UOXF

Slide | Video

Data Analytics and computer simulations for the study of cardiac disease
17:30 – 18:00 CompBioMed Task 3.2 meeting
Day 2 – Tuesday 17th March 2020
09:00 – 10:20 Organ Level modelling: imaging & combinations with simulation

Multiscale modelling & surrogate models, VVUQ etc

Ivan Benemerito, USFD

Slides | Video

Use of machine learning and modelling to improve diagnosis and treatment of ischaemic stroke
Robbie Sinclair, UCL

Sildes | Video

Generating Graphene Oxide structures with ML for Mechanical and Biological applications
Maxime Vassaux, UCL Integrating complexity in in silico models of biomaterials by means of concurrent multiscaling and automated data-base driven model reduction.
10:05 – 10:35 Break
10:35 – 11:35 HPC architectures en route to exascale

Middleware and Pilot Jobs

Peter Zinterhof, LRZ

Confidential

Big Data & AI in Medical Research and Life Science
Valeriu Codreanu, SURFsara

Slide | Video

High Performance Machine Learning at SURFsara
Andre Merzky, Rutgers

Slides | Video

RADICAL-Cybertools: Middleware Building Blocks for HPC & ML Workflows
11:35 – 13:00 Lunch
13:00 – 14:20 Data Management and Services

Training and Education

James Wilson, UCL

Slides | Video

UCL Institutional Infrastructure for FAIR Data
Stephan Hachinger, LRZ Research Data Management at LRZ
Narges Zarrabi, SURFsara

Slides | Video

Data Management in CompBioMed: Moving towards FAIR Data
Andrea Townsend-Nicholson, UCL

Slides | Video

Developing a new community of practice by engaging clinicians and biomedical scientists with advanced computational methods
14:20 – 14:50 Break
14:50 – 15:20 Virtual Round Table discussion
15:20 – 16:00 WP3 meeting

Register to attend here: https://www.eventbrite.co.uk/e/machine-learning-meets-modelling-and-simulation-methods-registration-90123589081