AI-Driven Materials Discovery: Machine Learning Potentials, Molecular Dynamics & High-Throughput Screening

The aim of this workshop is to provide researchers, academicians, and industry professionals with a practical understanding of how Artificial Intelligence and computational approaches are transforming materials discovery, enabling faster prediction, simulation, and optimization of next-generation materials.

Participants will learn to:

• Understand the fundamentals of AI-driven materials discovery and materials informatics workflows.
• Explore open-source materials databases and prepare datasets for AI-based analysis.
• Apply machine learning approaches for predicting material properties and performance.
• Understand machine learning interatomic potentials for accelerated atomic simulations.
• Learn molecular dynamics simulation concepts for studying material behaviour at atomic scales.
• Develop high-throughput screening workflows for automated material identification and ranking.
• Explore emerging AI approaches including generative AI, active learning, and autonomous materials discovery.

📅 Day 1: AI-Driven Materials Discovery & Machine Learning for Materials Informatics

• Focus: Understanding AI-based materials discovery workflows, materials data, and machine learning approaches for predicting material properties.
• Introduction to AI in materials science and its applications in energy materials, nanomaterials, catalysts, semiconductors, and advanced materials.
• Overview of materials informatics and open-source research databases:
  • Materials Project
  • OQMD
  • AFLOW
  • NOMAD
• Understanding materials representation:
  • Crystal structures
  • Chemical descriptors
  • Atomic features
  • Property-based datasets
• Introduction to machine learning models for materials prediction:
  • Regression models
  • Random Forest
  • Gradient Boosting
  • Neural Networks
  • Graph Neural Networks
• Developing AI workflows for material property prediction:
  Data Preparation → Feature Extraction → Model Training → Property Prediction
• Understanding model evaluation, validation strategies, and prediction accuracy metrics.

🛠️ Hands-on:

• Build an AI-based material property prediction model using open-source materials datasets.

🧰 Tools Covered: Google Colab, Python, Pymatgen, Matminer, Materials Project API, Scikit-learn, Pandas, Matplotlib

📅 Day 2: Machine Learning Potentials & Molecular Dynamics Simulation

• Focus: Learning AI-based atomic simulation methods and machine learning potentials for accelerated materials modelling.
• Introduction to molecular dynamics and atomic-level simulation approaches in materials research.
• Understanding Machine Learning Interatomic Potentials (MLIPs) and their role in replacing traditional force fields.
• Exploring advanced ML potential frameworks:
  • DeepMD
  • MACE
  • NequIP
  • Gaussian Approximation Potential (GAP)
• Understanding AI-based prediction of:
  • Atomic energies
  • Interatomic forces
  • Structural behaviour
• Molecular dynamics workflow:
  Structure Preparation → Energy Optimization → Simulation → Trajectory Analysis
• Applications of AI-assisted molecular dynamics:
  • Material stability
  • Diffusion behaviour
  • Thermal properties
  • Defect analysis

🛠️ Hands-on:

• Perform an AI-assisted molecular dynamics simulation using machine learning potentials.

🧰 Tools Covered: Google Colab, Python, ASE, MACE, ML Potentials, OVITO, Matplotlib

📅 Day 3: High-Throughput Screening & AI-Based Materials Discovery Pipeline

• Focus: Building automated AI-driven workflows for screening, ranking, and identifying promising materials candidates.
• Introduction to high-throughput materials screening and automated computational discovery workflows.
• Understanding AI-based material filtering and ranking strategies using large-scale materials databases.
• Property-based screening approaches:
  • Stability prediction
  • Formation energy analysis
  • Band gap prediction
  • Mechanical and thermal property evaluation
• Developing computational pipelines:
  Database → AI Prediction → Screening → Candidate Ranking
• Exploring AI optimization approaches:
  • Active learning
  • Generative AI for materials design
  • Multi-objective optimization
• Applications of AI materials discovery:
  • Battery materials
  • Catalysts
  • Hydrogen storage
  • Semiconductor materials
  • Nanomaterials

🛠️ Hands-on:

• Develop an AI-based high-throughput screening workflow to identify and rank potential material candidates.

🧰 Tools Covered: Google Colab, Python, Materials Project API, Pymatgen, Matminer, Scikit-learn, Plotly, Matplotlib

This workshop is designed for:

Researchers & PhD Scholars

• Materials Science
• Nanotechnology
• Computational Chemistry
• Physics
• Chemistry
• Chemical Engineering
• Energy Materials
• Computational Materials Modelling

Academicians & Faculty Members

• Researchers interested in integrating AI methodologies into materials research.
• Faculty members exploring computational approaches for teaching and research.

Industry Professionals

• R&D scientists working in:
  • Advanced materials development
  • Battery technologies
  • Semiconductor research
  • Catalysis
  • Nanomaterials
  • Materials informatics

Suitable For Participants Interested In:

• Artificial Intelligence in Materials Science
• Machine Learning for Scientific Research
• Molecular Simulation
• Computational Materials Discovery
• Data-Driven Material Design

After completing this workshop, participants will be able to:

• Understand the complete AI-powered materials discovery pipeline from data to prediction and validation.
• Work with open-access materials databases such as Materials Project, OQMD, and AFLOW.
• Build machine learning models for material property prediction.
• Understand and apply AI-based molecular simulation workflows.
• Gain practical knowledge of machine learning potentials and molecular dynamics approaches.
• Develop computational screening strategies for identifying promising materials.
• Apply AI-driven methodologies in research areas including energy materials, nanotechnology, catalysis, and advanced materials design.