AI for Materials Industry
A massive open online course focusing on connecting advanced AI techniques with invaluable materials expertise.
During ePotentia’s workshop on 27 June 2022 (10:00–17:00) you can gain hands-on experience in Materials AI by completing two practical case studies on the Flemish Supercomputer. Subscribe here and join us in Gent!
In the first case study, you will learn how to solve this problem using tabular data, as typically found in a spreadsheet such as Excel. You will learn how to create classical machine learning models (random forests), optimize them on the HPC and interpret them using explainable AI.
In the second case study the same problem will be treated with deep learning, more specifically convolutional neural networks. You will learn how to optimally train models on HPC GPU hardware, use data augmentation techniques to improve model quality and robustness, as well as regain insight using explainable AI.
A new way of manufacturing
The industrial materials enterprise is undergoing a revolutionary digital transformation where previously isolated machinery is being integrated to form interconnected networks capable of gathering vast amounts of information autonomously. Within this data, information lies hidden which would enable us to improve the quality of the produced materials and products, optimize the efficiency of production lines and even detect faults in equipment before they occur. To extract these invaluable insights, techniques are needed not only capable of finding this information, a proverbial needle in a haystack, but also of processing it and interpreting it in a way that allows us to implement the found insights without risk.
Artificial Intelligence (AI) focuses exactly on this type of data extraction, offering specialized techniques based on the type and amount of data. Many of these techniques are highly generic, enabling rapid transfer between different applications. Nonetheless, each application requires a careful combination of AI and materials expertise to fully exploit the its potential. Both new talent and existing experts will need to learn what Artificial Intelligence is and how it interacts with the intricacies of different parts of the materials enterprise. While many courses already exist on either Artificial Intelligence, Manufacturing and materials science, none currently specialize in exploring the connection between them.
Artificial Intelligence for Materials Industry (AI4MI) aims to fill this niche by introducing Artificial Intelligence through the perspective of carefully selected industrial use cases in the form of a massive open online course (MOOC). In this way, the materials expert learns to integrate AI in their production lines and R&D projects, while the AI professional learns to watch out for the specific pitfalls related to materials industry projects. Education can happen at the users own pace, and easily returning to it in the future whenever necessary. AI4MI aims to provide the knowledge necessary to interconnect not only the machines, but also the people needed to maintain a competitive advantage in Industry 4.0s materials manufacturing.