From the Theatre: How AI is redesigning the battery - with the Faraday Institution
Summary
This episode from the RI Science Podcast features three experts discussing how AI is transforming battery science across different scales. Aaron Walsh explains how machine learning navigates the vast atomic space to design new materials, while Mona explores using AI to predict battery degradation and remaining useful life at the system level. Sam Cooper details how computer vision and generative models optimize microstructure for better performance. Together, they highlight how AI accelerates material discovery, improves device longevity, and streamlines manufacturing processes for next-gen…
Topics discussed
Introduction: The role of AI in battery research
Historical context: From atoms to the periodic table
Machine learning and the vast space of material combinations
Challenges in identifying next-generation battery materials
Generative AI and text-to-material models like Chameleon
AI co-scientists and autonomous research workflows
Scaling AI for industrial problems and Cusp AI
Speaker introduction: Background in control engineering
Factors causing battery degradation and capacity loss
The scale of battery demand and data complexity
AI models for predicting remaining useful life
Implementing AI models in real-world hardware constraints
Narrow AI and the intersection of data, power, and model
Microstructure: Internal structure and ion transport
Computer vision for material segmentation and mapping
Generative AI for creating 3D electrode structures
Robotic labs and AI-driven experimental validation
Listen ad-free on Castria