Talks
Invited and contributed talks given by DDMMS members at conferences, workshops and seminars.
-
2026
ML-PEG: Machine Learning Performance and Extrapolation Guide
Contributed talkSession 1 (9–11 September 2026). Presentation on ML-PEG, evaluating machine learning interatomic potential performance and extrapolation boundaries.
-
2026
ML-PEG: Machine Learning Performance and Extrapolation Guide
Invited talkSession: AI agents and Autonomous Labs (3–4 September 2026 at Max Planck Institute for Polymer Research).
-
2026
MMM Hub Software Spotlight: Janus ML
Invited talkInvited presentation (26 August 2026). Overview of the Janus ML software suite, practical demonstration of job submission and workflows on the MMM Hub Young Tier 2 supercomputer, and applications in machine-learned interatomic potentials from recent research.
-
2026
Bridging the Gap Between Theory and Experiment: Machine Learning Inter-atomic Potentials for Metal Organic Frameworks
Contributed talkSession 3: Bulk (Thursday 2 July 2026). Presenting advances in machine learning interatomic potentials for metal-organic frameworks, connecting theoretical vibrational properties with experimental characterisation.
-
2026
ML-PEG: Machine Learning Performance and Extrapolation Guide
Contributed talkSession 1: Algorithms (Wednesday 1 July 2026). Presentation on ML-PEG, co-authored with Joseph Hart, Ilyes Batatia, Domantas Kuryla, Alin M. Elena, and Gábor Csányi, introducing rigorous benchmarks and interactive evaluation for foundation MLIPs.
-
2026
Bridging the Gap Between Theory and Experiment: Augmenting MD-EXAFS with Machine Learning Interatomic Potentials for Complex Materials Research
Contributed talkSession: Computational Methods and Data Analysis (21–26 June 2026, Chiang Mai, Thailand).
-
2026
Machine Learned Potentials for Phonon Calculations of Metal–Organic Frameworks
Contributed TalkHigh-throughput phonon calculations in metal-organic frameworks using fine-tuned MACE foundation models within the quasi-harmonic approximation.
-
2026
Machine Learning Performance and Extrapolation Guide (ML-PEG)
Invited talkInvited talk presented with James Kermode in the session on Infrastructure Developments. Covers ML-PEG, a modular framework to develop, run, and visualise automated, interactive test suites for machine-learned interatomic potentials.
-
2025
Machine Learnt Interatomic Potentials for High-Throughput Phonon Calculations of Metal-Organic Frameworks
Invited talkInvited talk (22 August 2025, Realfagbygget) at the COSY (COmputational multi-Scale materials societY) Gemini Center event in collaboration with the Friday Physics Colloquium. Presents machine-learned potentials for high-throughput phonon calculations and thermodynamic property prediction in metal-organic frameworks.
-
2025
Keynote: HPC Best Practices
Keynote talkOpening keynote lecture (3 July 2025) delivered at the ICTP Magurele Summer School (smr 4077), addressing modern high-performance computing best practices, parallel optimization principles, and scalable scientific simulation workflows.
-
2025
Materials Highlight: Machine Learned Interatomic Potential for High-Throughput Phonon Calculations of Metal-Organic Frameworks
Showcase talkMaterials Highlight talk (19 June 2025, RAL Visitor Centre). Presenting MACE-MP-MOF0, a fine-tuned machine learning potential and high-throughput workflow that corrects unphysical imaginary modes and enables accurate thermal expansion and mechanical property predictions across diverse MOF structures.
-
2025
Machine Learned Interatomic Potentials for Modelling Porous Materials
Invited talkInvited talk (15 May 2025) in Session 3: Recent Advances in Computational Materials Science, hosted by the National Institute of Materials Physics (NIMP). Explores fine-tuned machine learning potentials (MACE-MP-MOF0) for high-throughput phonon and thermodynamic calculations in zeolites and metal-organic frameworks.