Conference Logo ProgramMaster Logo
Conference Tools for 2027 TMS Annual Meeting & Exhibition
Login
Register as a New User
Help
Submit An Abstract
Propose A Symposium
Presenter/Author Tools
Organizer/Editor Tools

About this Abstract
Meeting 2027 TMS Annual Meeting & Exhibition
Symposium Algorithms Development in Materials Science and Engineering
Presentation Title Model-Agnostic Thermodynamic Sampling
Author(s) Thomas D. Swinburne, Mihai-Cosmin Marinica
On-Site Speaker (Planned) Thomas D. Swinburne
Abstract Scope Prediction of bulk / defect phase diagrams is a central task of atomic simulation, but thermodynamic sampling of configuration space is extremely costly as high degrees of convergence are required. I will discuss recent work that provides a new scheme to compute free energies, where the atomic energy model is not precisely defined a priori. Instead, score matching is employed to learn a Shannon entropy in the latent space common to most modern machine learning interatomic potentials, from which the Gibbs free energy of a particular energy model can be predicted instantaneously by taking the convex (Fenchel-Lagrange) conjugate. The approach enables two new capabilities: uncertainty on energy model parameters can be instantly propagated to phase diagrams and experimental data can be used to fine-tune interaction potentials. Nat Comm 2026 https://doi.org/10.1038/s41467-025-66938-8
Proceedings Inclusion? Planned:
Keywords Machine Learning, Phase Transformations, Modeling and Simulation

OTHER PAPERS PLANNED FOR THIS SYMPOSIUM

A concurrent atomistic-continuum method for coupled dislocation-fracture simulations
A Finite-Temperature Atomistic Framework for Mass Transport in Metallic Alloys
A Multiple Time Scale Methodology for Predictive Simulation of Finite-Temperature Materials Processes
A No-Code Graphical Framework for Forward and Inverse Physics-Informed Neural Networks
A Parameterized Simulation Framework for Life Prediction and Design of High-Temperature Components
A simulated annealing technique for atomic environment rectification
A Static Scheduling Method for Steelmaking–Continuous Casting Considering BOF–VD Steam Generation and Consumption Balance
A Validated ML-Potential Model to Quantum Algorithms Workflow for Constraint-Aware Defect-Configuration Search in Rare Earth Elements Doped ZrO₂
Accelerating phase field simulations of microstructure evolution in additive manufacturing via tensor decomposition
Adaptive Knowledge Discovery for Alloy Spall Strength through Self-Evolving Knowledge Graphs
AI-Assisted Evaluation of Deformation During In Situ Micropillar Compression of Irradiated Ferritic/Martensitic Steels
An Agentic Workflow for Autonomous Density Functional Theory Simulations and Data-Driven Materials Discovery
An Efficient Approach to Path Planning in a High Dimensional Thermo-Spatial Material System for Optimal Gradient Materials
An improved phase-field model for grain growth with arbitrary misorientation-dependence of grain boundary energy
An Open Source Continuous Orientation Distribution Algorithm in Python
Automated Microstructure Reconstruction, Mesh Generation, and Deep Learning Algorithms for Modeling Materials
Autoregressive latent diffusion for 3D polycrystalline material dynamics
BESFEM: Battery Electrode Simulation using Finite Element Method
BlobNet: A New Lattice-Agnostic Tool for Angstrom Resolution Atom Finding
Capacity Planning and Resource Configuration Optimization for Multiple Production Lines in Process Manufacturing
Chemical Basis Sets and Average-atom Potentials for Disordered Multicomponent Alloys
Closing the Loop from Experiment to Simulation: A FAIR, Ontology-Based Platform for Dynamic Materials Testing
Contract-Driven Production Planning for Steelmaking–Continuous Casting Based on Resource Allocation Optimization
Data-Driven Fail-Safe and Safe-Life Design for Non-Uniform Lattice Structures
Data driven discovery of phase field crystal models from atomistic simulations
DD2MD: An Exact Geometric Algorithm for Seeding Molecular Dynamics with Prescribed Dislocation Microstructures
Deep Learning on DFT-Derived Charge Densities for Stacking Fault Energy Predictions in High Entropy Alloys
Detailed Analysis of Ising-Model-Based Molecular Structure Searches in Four-Atom Systems and Expansion to Other Systems
Driving Force Extension for Stable Multi-Order Parameter Phase-Field Models
Emergence of Quality Fluctuations in Combined Machining and Heat-Treatment Processes: A Multi-Agent Simulation Approach
FFT-based modelling with direct input from microstructural images
Fitting semi-empirical interatomic potentials to large ab initio databases
Fracture-Aware Topology Optimization With and Without Gradients on a GPU-Native Finite-Element Framework
Implementation of Bayesian Optimization over Function Networks with Partial Evaluations for Cost-Aware AI Guided Material Discovery of Critical Mineral Lean Aluminum Alloys
Learning trustworthy equations from materials data with neural networks
Low-Rank Tensor Completion for Multi-Property Prediction of Polymers
Microstructure-based modeling of stress corrosion cracking at a crack tip
MISalign: Human-in-the-loop Image Alignment and Blending for Communication and Computation
Model-Agnostic Thermodynamic Sampling
Novel framework to quantify helium bubble pressure in irradiated copper via nanoscale strain mapping and physics-based modeling
Optimized Design of Steel Planar Frames Using Genetic Algorithms and Semi-Rigid Connection Databases
PINN Approach to Accelerated Spin Dynamics Simulation
Prediction of Uniaxial Fatigue, Non-Uniaxial Fatigue and Their Variabilities Through Microstructure-Based Elasto-Viscoplasticity Modeling
Quantification of microstructure variability with sparse data
Quantification of microstructures in polycrystalline materials from atomistic calculations
Quantifying the Impact of EBSD Segmentation Workflows on Grain Structure Metrics
Quantitative Phase-Field Benchmarking of Eutectic Fraction in Laser-Deposited Hypoeutectic Al–Si
Rapid Part-Scale Temperature Prediction for Additive Manufacturing via Bayesian-Calibrated Physics-Based Modeling
Rate Modeling of Thermally Activated Defect Evolution in BCC Fe using Transition State Theory
Simulation-Informed Deep Learning for Morphology-Based Parameter Estimation in Fe–Cr Alloys
Surrogate modeling of creep in nuclear materials using mixed adaptive basis functions with spatially varying parameters
Temporal Acceleration of Molecular Dynamics Simulations for Iron using a Parareal Framework
The Lattice Parameter and its Role in Alloy Development
Questions about ProgramMaster? Contact programming@programmaster.org | TMS Privacy Policy | Accessibility Statement