Seminars
WCPM hosts regular seminars where we invite speakers working both in fundamental methodology and on applications of predictive modelling and UQ.
From October 2016, the seminars are on Mondays, 1pm and joint with the Centre for Scientific Computing, with a focus on all aspects of research involving significant computational resource. Venues are alternating between PS0.17a (Physical Sciences) and D2.02 (Engineering), and a sandwich lunch is available from 12:45pm.
If you would like to be kept informed of upcoming seminars, please do sign up here or send an email to wcpm-seminar-join at listserv dot csv dot warwick dot ac dot uk.
Past presentations are available below.
Upcoming Seminars
Ozsel Kilinc (WMG)
Multi-agent Deep Reinforcement Learning with Extremely Noisy Observations
Reinforcement Learning (RL) is concerned with enabling agents, i.e. decision makers, to learn how to accomplish a task by taking sequential actions in a given stochastic environment and relies on Markov Decision Processes (MDP). In recent years, Deep Reinforcement Learning (DRL), which leverages RL approaches using Deep Neural Network based function approximators, has been proved to achieve human-level performance in a number of applications requiring an individual agent, mostly gaming environments. Many real-world applications, on the other hand, can be modelled as multi-agent systems, e.g. autonomous driving, energy management, fleet control, trajectory planning, and network packet routing. Straightforward applications of single-agent DRL methodologies for learning a multi-agent policy are not well suited due to severe non-stationarity and partial observability issues inherently existing in multi-agent systems. In this talk, we will first discuss these problems and briefly introduce the existing multi-agent RL (MARL) approaches in the literature. Then, we will consider a particularly challenging multi-agent learning problem whereby most agents' observations are both partial and extremely noisy, hence the existing MARL techniques fail to learn an optimal policy. We will present a multi-agent deep deterministic policy gradient algorithm enhanced by a communication medium (MADDPG-M) which implements a two-level, concurrent learning mechanism to overcome the extremely partial observability and the accompanying non-stationarity. In this model, an agent's policy depends on its own private observations as well as those explicitly shared by others through a communication medium. As we consider a medium with limited capacity, at any given point in time, each agent must decide whether its own observations are sufficiently informative to be shared with others. Our experimental results demonstrate that the algorithm performs well in six highly non-stationary environments of progressively higher complexity, and offers substantial performance gains compared to the existing MARL baselines in the literature.
Past Seminars (2018/19)
| Date | Speaker | Presentation |
| 11 March 2019 |
Fritz-Haber Institute, Berlin |
Nuclear quantum effects in weakly bonded systems
|
| 4 March 2019 |
Warwick Medical School |
The light microscopy imaging pipeline at Warwick Medical School
|
| 25 February 2019 |
Institute for Risk and Uncertainty, Liverpool |
Connections between Rare-Event Simulation and Bayesian Inference
|
| 18 February 2019 |
|
Phase field methods for simulating ferroelectrics and other materials
|
| 11 February 2019 |
CEA Saclay, France |
Free energy landscape of point defects in body centered cubic metals |
| 4 February 2019 |
Warwick Manufacturing Group |
Multi-agent Deep Reinforcement Learning with Extremely Noisy Observations
|
| 28 January 2019 |
Babraham Institute, Cambridge |
Modelling of Chromosome Structure and Function
|
| 21 January 2019 |
Warwick Mathematics Institute / Turing Institute |
Predicting risk of hospital admission and readmission of Scottish population; Measuring Sample Quality with diffusions
|
| 14 January 2019 |
Elementary Particle Physics Group, Warwick Physics |
Reconstruction of neutrino-induced events in Liquid-Argon Time-Projection Chamber (LArTPC) Detectors
|
| 3 December 2018 |
Institute for the Physics of Living Systems, University College London |
Minimal coarse-grained models for biological assembly
|
| 26 November 2018 |
School of Engineering, University of Warwick |
Stochastic Simulation of Reaction-Diffusion Systems
|
| 19 November 2018 |
Faculty of Engineering, Department of Materials, Imperial College |
Simulating charged defects in 2d materials: how we taught electronic screening to machines
|
| 12 November 2018 |
Warwick Centre for Predictive Modelling, School of Engineering, University of Warwick |
Modifying the bandstructure of half-Heuslers to improve their thermoelectric power factor
|
| 05 November 2018 |
Faculty Research Centre in Flow Measurement and Fluid Mechanics, Coventry University |
Monte Carlo methods for massively parallel architectures
|
| 29 October 2018 |
Warwick Centre for Predictive Modelling, University of Warwick |
Electron-spin mediated distortion in metallic systems
|
| 15 October 2018 |
Warwick Centre for Predictive Modelling, University of Warwick |
First principles design of new multiferroic materials
|
| 08 October 2018 |
Andrew Davis UK Atomic Energy Authority, Culham Centre for Fusion Energy, Abingdon |
High Performance and CAD based Radiation Transport
|
Anđela Šarić
Adam Noel
Johannes Lischner
Chatu Kumarasinghe
Martin Weigel
Gautam Anand
Berk OnatPast Seminars (2017/18)
| Date | Speaker | Presentation |
| 25 June 2018 |
School of Mathematics and Physics |
Variations on kernel density estimation |
| 18 June 2018 |
Department of Mathematical Sciences |
A performance portable framework for molecular simulations |
| 11 June 2018 |
Physics Department |
A predilection for precise prediction: trying to model materials reliably |
|
21 May 2018 |
School of Biological and Chemical Sciences, Queen Mary University of London (QMUL) |
Modelling aggregation-induced emission in organic crystals |
| 16 May 2018 |
University of Texas at Austin |
Correlating structure and function for nanoparticle catalysts Abstract Slides |
| 30 Apr 2018 |
Chemical and Process Engineering |
Multiscale modelling of surfaces and interfaces |
| 30 Apr 2018 |
Warwick Mathematics Institute |
Predicting Rare Events via Large Deviations Theory: Rogue Waves and Motile Bacteria Abstract Slides |
| 12 Mar 2018 |
Warwick Centre for Predictive Modelling, School of Engineering, University of Warwick |
Large-scale real-space electronic structure calculations Abstract Slides |
| 05 Mar 2018 |
Department of Statistics, University of Oxford |
Biologically inspired de novo protein structure prediction Abstract Slides |
| 26 Feb 2018 |
Mechanical Engineering, University of Michigan, Ann Arbor, USA |
Large-scale real-space electronic structure calculations Abstract Slides |
| 19 Feb 2018 |
Tribology and Wear, Fraunhofer IWM, Freiburg |
Tribology of diamond and silicon: atomic-scale insights from computers simulations Abstract Slides |
| 12 Feb 2018 |
Department of Aeronautics, Faculty of Engineering, Imperial College London |
Uncertainty Quantification and Aircraft Engines Abstract Slides |
| 05 Feb 2018 |
Department of Physics and Astronomy, University of Sheffield |
Unusual Dynamic Processes in Polymer Nanocomposites Abstract Slides |
| 29 Jan 2018 |
Warwick Centre for Predictive Modelling, School of Engineering, University of Warwick |
Transport in two-dimensional materials Abstract Slides |
| 22 Jan 2018 |
Met Office |
Modelling clouds for weather, climate and beyond… Abstract Slides |
| 15 Jan 2018 |
Faculty of Engineering, University of Bristol |
Multi-scale cellular automata finite element (CAFE) framework for fracture in heterogeneous materials: to exascale via coarrays/MPI hybrid programming Abstract Slides |
| 4 Dec 2017 |
Ben Chapman Demetris Marnerides |
Ion cyclotron emission as a diagnostic of the time evolution Deep Learning for the High Dynamic Range Imaging Pipeline |
| 27 Nov 2017 |  Magnus RichardsonZeeman Mathematical Biology Institute, Warwick |
Neurons, Synapses and neuromodulators |
| 23 Nov 2017 |
|
DFTB+, the fast way of doing quantum mechanics Abstract Slides |
| 06 Nov 2017 |  Felix HofmannDepartment of Engineering Science University of Oxford |
Atomic Scale Defects: Probing Structure and Function Abstract Slides |
| 30 Oct 2017 |
Laura Ratcliff, Imperial College Jolyon Aarons, Warwick Physics Luigi Genovese, L_Sim, CEA Grenoble |
Upscaling DFT Capabilities for the Many Thousand Atoms Regime Metallic DFT On Nanoparticles with Thousands of Atoms Opportunities from Accurate and Efficient Density Functional Theory |
| 23 Oct 2017 |  Juergen BrankeWarwick Business School |
|
| 16 Oct 2017 |  Reinhard MaurerDepartment of Chemistry, University of Warwick |
Controlled Self-Assembly, Switching, and Catalysis: |
| 9 Oct 2017 |  Gabriele SossoDepartment of Chemistry, Unviersity of Warwick |
Predicting Crystal Nucleation Rates via Atomistic Simulations: The Case of Ice Nucleation Abstract Slides |
Graeme Henkelman
Karen Johnston
Laura de Sousa Oliveira
Charlotte Deane
Vikram Gavini
Gianpietro Moras
Francesco Montomoli
Nigel Clarke
Vasileios Vargiamidis
Ian Boutle
Anton Shterenlikht
Bálint AradiPast Seminars (WCPM & CSC, 2016/17)
| Date | Speaker | Presentation |
| 26 June 2017 |
WCPM & School of Engineering, University of Warwick |
Modular Bayesian uncertainty assessment for structural health monitoring
|
|
WCPM & School of Engineering, University of Warwick |
Quantum transport simulations for understanding the thermoelectric effect in nanocomposites
|
|
| 19 June 2017 |
Research Software Engineering, University of Sheffield |
Bad software and how it's ruining your research
|
| 12 June 2017 |
Mathematics Institute and Center for Complexity Science, University of Warwick |
Oscillatory kinetics in cluster-cluster aggregation
|
| 05 June 2017 |
Department of Chemistry, University of Memphis |
The role of transition metals in the structure and reactivity of astrochemicals
|
| 22 May 2017 |
Engineering and Materials Science, Queen Mary University of London |
Cracking me softly – the mechanics of hyperelastic Kirigami structures
|
| 15 May 2017 |
Physical Sciences, University of Kent |
First principles design of new multiferroic materials
|
| 13 March 2017 |
Warwick Life Sciences |
Design principles of cellular systems from evolutionary and systems dynamics perspectives |
| 6 March 2017 |
Reading Chemistry |
Computational thermodynamics: how to calculate phase diagrams without the fuss |
| 27 February 2017 |
Karlsruhe Institute of Technology |
Contact and cavitation: Computer models for tribological processes |
| 20 February 2017 |
Warwick Physics |
The dynamics and flexibility of biomedically important proteins: a combined computational/experimental approach to tackle antimicrobial resistance (AMR)
|
| 13 February 2017 |
Warwick Maths / WMG |
Computational and mathematical modelling of acoustic linears in aircraft engines
|
| 6 February 2017 |
Head of Research Software Development, UCL |
Computational Science As A Service: A new institutional model for research software engineering
|
| 30 January 2017 |
University of Bristol |
What is good research software, how can it be engineered, and is it deliverable within the constraints of academic projects?
|
| 23 January 2017 |
Connected Systems Research Group, School of Engineering, Warwick |
Exploring the Potential for FPGA Scientific Computing Research at Warwick
|
| 16 January 2017 |
|
Interfacial barriers to thermal transport in carbon nanomaterials and polymer-nanotube composites: a molecular modelling study
|
| 5 December 2016 |
|
Diabatization strategies for non-adiabatic dynamics: photodissociation of acetylene and NO/Au 3 scattering
Abstract  |
| 28 November 2016 |
|
|
| 21 November 2016 |
|
|
| 14 November 2016 |
|
|
| 07 November 2016 |
|
|
| 31 October 2016 |
|
|
| 24 October 2016 |
|
|
| 17 October 2016 |  Tom HudsonMathematics Institute, University of Warwick |
André Jesus
Samuel Foster
Mike Croucher
Colm Connaughton
Nathan DeYonker
Ettore Barbieri
Nicholas Bristowe
Orkun Soyer
Lars Pastewka
James Elliot
Christopher Robertson
Mikhail Poluektov
Past Seminars 2015/2016
| Date | Speaker | Presentation |
| 31st May 2016 |  Ben GrahamDepartment of Statistics, University of Warwick |
|
| 24th May 2016 |   Duncan Lockerby/David StephensonSchool of Engineering, University of Warwick |
Accelerating a multiscale continuum-particle fluid dynamics model
|
| 17th May 2016 |
 Gareth Conduit Department of Physics, University of Cambridge |
|
| 10th May 2016 |
 Torgyn Shaikhina School of Engineering, University of Warwick |
Machine Learning for predictive modelling based on small biomedical and clinical data
Abstract (PDF), Slides (PDF) |
| 3rd May 2016 |
Department of Statistics, University of Warwick
|
Bayesian inference and model selection for stochastic
epidemics and other coupled hidden Markov models
(with special attention to epidemics of
|
| 26th Apr 2016 |
Bond-selective chemical reactivity from first principles:
|
|
| 8th Mar 2016 |
Predictive Multi-scale Modelling - an Interdisciplinary
Perspective and the Case of Density Embedding Methods
|
|
| 1st Mar 2016 |
Department of Physics, University of Warwick
|
Quantum Engineering for electrons and spins
|
| 16th Feb 2016 |
Dynamics of Large Fluctuations: from Chaotic Attractors to Ion Channels |
|
| 9th Feb 2016 |
University of Warwick
|
Image clustering based on camera fingerprints |
| 2nd Feb 2016 |  Scott HabershonDepartment of Chemistry, University of Warwick |
Progress and perils in predictive computational catalysis Abstract (PDF) |
| 26th Jan 2016 |
 Charles Gadd Warwick Centre for Predictive Modelling, University of Warwick |
Uncertainty Quantification with Gaussian Process Latent
Variable Models |
| 19th Jan 2016 |
Warwick Centre for Predictive Modelling,
University of Warwick |
Uncertainty Quantification Using Deep Gaussian Processes
and Variational Bayesian inference
|
| 8th Dec 2015 |
School of Mathematical Sciences, University of Nottingham
|
Novel stochastic thermostats for rigid body dynamics
|
| 26th Nov 2015 |
Technische Universität München |
Predictive Coarse-Graining Abstract (PDF), Slides (PDF) |
| 17th Nov 2015 |
Department of Physics, University of Cambridge
|
Computational Materials Discovery meets Experiment Abstract (PDF) |
| 10th Nov 2015 |
|
Towards a multiscale framework for robust simulation of inelastic dynamic processes Abstract (PDF), Slides (PDF) |
| 3rd Nov 2015 |
|
Reliable Knowledge: How far can we trust electronic structure simulations? Abstract (PDF) |
| 27th Oct 2015 |
|
Exchange-Correlation Functional with Uncertainty
Quantification Capabilities for Density Functional Theory |
| 20th Oct 2015 |
|
Pore-scale analysis of dynamics of two-phase flow
in porous media |
| 13th Oct 2015 |
|
Dimensionality Reduction Methods in Predictive Modelling |
| 13th Oct 2015 |
|
Approximate Bayesian Inference for Machine Learning |
| 6th Oct 2015 |
|
On the predictive capabilities of multiphase Darcy flow |
Simon Spencer
Rudolf Römer
Alireza Daneshkhah
Michael V. Tretyakov
Markus Schöberl
Andrew Morris
Savvas Triantafyllou
Keith Refson
Manuel Aldegunde
Stela Makri
Louis Ellam
Matteo IcardiPast Seminars 2014/2015
| Date | Speaker | Presentation |
| 25th Jun 2015 |
|
Excited state calculations and theoretical spectroscopy of complex nanomaterials using Linear-Scaling Density Functional Theory Abstract (PDF), Slides (PDF) |
| 18th Jun 2015 |
|
Probabilistic Numerical Methods for Deterministic Differential Equations Abstract (PDF), Slides (PDF) |
| 11th Jun 2015 |
|
Applying machine learning to quantum mechanics Abstract (PDF), Slides (PDF) |
| 4th Jun 2015 |
|
Sampling constrained probability distributions using Spherical Augmentation Abstract (PDF), Slides (PDF) |
| 28th May 2015 |
|
Computational modelling for performance improvement of polymer nanocomposites Abstract (PDF), Slides (PDF) |
| 15th May 2015 |
|
Uncertainties in classical effective potentials: sources and quantification strategies Abstract (PDF), Slides (PDF) |
| 7th May 2015 |
|
Predictive Modelling: a view from the atomic level Abstract (PDF), Slides (PDF, PPTX) |
| 30th Apr 2015 |
|
Modelling crystal nucleation and growth Abstract (PDF), Slides (PDF, PPTX) |
| 23rd Apr 2015 |
|
Scaling Limits in Computational Bayesian Inversion Abstract (PDF), Slides (PDF) |
| 12th Mar 2015 |
|
Density functional theory and slowly varying fluctuations at finite temperature to describe alloy and magnetic phase diagrams Abstract (PDF), Slides (PDF) |
| 5th Mar 2015 |
|
Molecular Dynamics with on-the-fly Machine Learning of Quantum |
| 26th Feb 2015 |
|
Differential Geometric MCMC Methods |
| 19th Feb 2015 |
|
Efficient emulation of high-dimensional outputs using manifold |
| 12th Feb 2015 |
|
Electronic, thermal, and thermoelectric transport in nanostructures |
| 5th Feb 2015 |
|
Quasi- and Multilevel Monte Carlo Methods for Bayesian Inverse |
| 29th Jan 2015 |
|
Constructing and sampling graphs with specified joint-degree matrix |
| 22nd Jan 2015 |
|
Predictive Modelling of Riboregulatory Circuits to Re-engineer Living Cells |
| 15th Jan 2015 |
Tim Sullivan |
Consistency, Robustness, and Brittleness of Bayesian Inference |
| 8th Jan 2015 |  Jesper KristensenApplied and Engineering Physics, Cornell University, Ithaca NY |
Uncertainty Quantfication with Surrogate Models in Alloy Modelling |
Nicholas Hine
Andrew Stuart
Gábor Csányi
Shiwei Lan
Łukasz Figiel
Peter Brommer
Mark Rodger
Claudia Schillings
Julie Staunton
James Kermode
Neophytos Neophytou
Aretha Teckentrup
Charo del Genio
Alfonso Jaramillo