AI-Driven Catalyst Discovery: High-Throughput Screening, Selectivity Prediction & Sustainable Process Optimization

To equip researchers and industry professionals with practical skills in AI-driven catalyst screening, catalytic performance prediction, and explainable machine learning for efficient catalyst selection and sustainable reaction optimization.

• Understand AI-driven catalyst discovery and high-throughput computational screening.
• Analyze catalyst descriptors and structure–performance relationships.
• Process and visualize experimental catalytic performance datasets.
• Develop machine learning models for catalytic activity and selectivity prediction.
• Evaluate predictive models using statistical performance metrics.
• Interpret influential catalyst properties and reaction parameters using SHAP.
• Apply constrained optimization to identify promising reaction conditions.

📅 Day 1: Catalyst Informatics & High-Throughput Screening

• AI-driven approaches to heterogeneous catalyst discovery.
• Experimental catalyst datasets and catalytic performance indicators.
• Catalyst descriptors: composition, metal loading and surface properties.
• Structure–activity and structure–selectivity relationships.
• Data preprocessing, normalization and descriptor engineering.
• High-throughput catalyst filtering and performance-based ranking.
• Catalyst shortlist generation and comparative visualization.

🛠️ Hands-On: Data-Driven Catalyst Screening & Ranking

• Activity: Clean and analyze catalyst datasets, visualize key descriptors and rank promising candidates.
• Tools: Google Colab, Python, Pandas, Seaborn.
• Output: Cleaned dataset, descriptor plots and ranked catalyst shortlist.

📅 Day 2: Machine Learning for Catalytic Activity & Selectivity Prediction

• Machine learning workflows for catalytic performance prediction.
• Feature selection for catalyst descriptors and reaction parameters.
• Catalytic activity prediction using Random Forest regression.
• Reaction selectivity prediction using XGBoost.
• Model training, cross-validation and hyperparameter tuning.
• Predictive performance evaluation using R², MAE and RMSE.
• Predicted-versus-experimental performance visualization.

🛠️ Hands-On: ML-Based Catalyst Performance Prediction

• Activity: Train and evaluate activity and selectivity prediction models using catalyst experimental data.
• Tools: Google Colab, Python, Scikit-learn, XGBoost.
• Output: Trained predictive models, evaluation metrics and prediction plots.

📅 Day 3: Explainable AI & Sustainable Reaction Optimization

• Explainable AI for interpreting catalytic performance models.
• SHAP-based feature importance and catalyst descriptor analysis.
• Identification of key activity- and selectivity-driving parameters.
• Reaction temperature and operating-condition optimization.
• Multi-objective optimization of activity and selectivity.
• Sustainability-aware screening using energy-related operating proxies.
• AI-guided catalyst selection and experimental validation strategies.

🛠️ Hands-On: SHAP-Guided Reaction Optimization

• Activity: Interpret model predictions using SHAP and optimize reaction conditions within defined operating constraints.
• Tools: Google Colab, Python, SHAP, XGBoost, SciPy.
• Output: SHAP plots, optimized reaction parameters and comparative performance results.

This workshop is designed for researchers, academicians, postgraduate students, and industry professionals working in:

• Catalysis & Chemical Sciences: Heterogeneous catalysis, catalyst design, reaction kinetics, and catalytic reaction engineering.
• Materials Science & Nanotechnology: Functional materials, nanocatalysts, surface engineering, and catalyst characterization.
• Chemical & Process Engineering: Chemical process optimization, industrial catalysis, and reaction engineering.
• Green Chemistry & Sustainability: Sustainable catalysis, resource-efficient chemical processes, and cleaner production technologies.
• Artificial Intelligence & Computational Research: Machine learning, materials informatics, predictive modeling, and data-driven catalyst discovery.
• Industrial R&D: Catalyst development, chemical manufacturing, process improvement, and sustainable technology development.

Recommended Background: Basic understanding of chemistry, catalysis, or chemical engineering. Prior Python or machine learning experience is helpful but not mandatory.

Upon successful completion of this workshop, participants will be able to:

• Screen Catalyst Candidates: Prepare experimental datasets and identify promising catalysts through descriptor-based analysis.
• Build Predictive Models: Develop Random Forest and XGBoost models for catalytic activity and reaction selectivity.
• Evaluate Model Performance: Assess prediction accuracy using R², MAE, RMSE, and visualization techniques.
• Interpret AI Predictions: Apply SHAP to identify influential catalyst descriptors and operating parameters.
• Optimize Reaction Conditions: Use computational optimization to recommend reaction conditions under defined performance and operating constraints.
• Support Research Decisions: Integrate screening, predictive modeling, and explainable AI into catalyst research workflows.