Explore the intersection of machine learning and physics in materials innovation through this comprehensive 48-minute webinar. Discover how Schrödinger's tools address challenges in data-driven AI-based design for materials science and chemistry. Learn about using molecular dynamics simulations to enhance ML model accuracy for viscosity predictions, building interpretable models for Li-ion battery electrolyte ionic conductivity, and improving ML performance for organic electronics properties using density functional theory. Gain insights into AutoQSAR, DeepAutoQSAR, chemical featurization, optoelectronics properties prediction, volatility machine learning, and neural network potentials. Understand the applications of these techniques in diverse fields such as redox flow batteries, transparent conducting oxides, and liquid electrolytes. Explore Schrödinger's LiveDesign® informatics platform and its suitability for various materials and data types.
Data-Driven Materials Innovation: Where Machine Learning Meets Physics
Overview
Syllabus
Data-driven materials innovation: where machine learning meets physics
Machine Learning for Materials Design/Discovery at Schrödinger
Supervised Learning in Materials Science
Featurization in Diverse Materials Systems
Automated Machine Learning and Visualization in Molecular Systems
AutoQSAR for Ionic Liquids
DeepAutoQSAR: Automated Model Selection & Parameter Optimization
Case Study - Redox Flow Batteries
AutoQSAR vs DeepAutoQSAR Results
Chemical Featurization using Physics
Customized Polymer Descriptors Outperform Simple Monomers
Viscosity Dataset for Machine Learning Module
Quantitave Structure-Property Relationships QSPR
Impact of MD-Derived Simulation Descriptors
Impact of MD-Derived Simulation Descriptors
Machine Learning Optoelectronics Properties with DFT descriptors
Database of Optical Properties of Organic Compounds
Benchmark of DFT Descriptors
Feature Importance Analysis
Machine Learning for Volatility of Organic Molecules
Evaporation/Sublimation of Organic Molecules
Benchmarking ML Algorithms
Prediction of Pressure-Temperature Relationships
Applications of Volatility Machine Learning
Machine Learning for Inorganic 3D Crystal Structures
Transparent Conducting Oxide Band Gap ML
User Interface
DeepAutoQSAR Results
Machine Learning Property Prediction Panel
ML for Formulations
Active Learning and Genetic Optimization
Active Learning OptoElectronics Multi-Parameter Optimization MPO
Active Learning Workflow for OptoElectronics
Optoelectronic Genetic Optimization
Machine Learning Forcefields
Neural Network Potentials NNPs
Our First NN Model: Schrödinger-ANI SANI
QRNN: Charge-Recursive Neural Network
Bulk Properties of Liquid Electrolytes
Enterprise Informatics
Schrodinger's Informatics Platform - LiveDesign®
Suitable for Diverse Materials and Data Types
Summary
Thank you
Taught by
nanohubtechtalks
