Second Focus Group shares machine learning experiences
Intellegens hosted the second Alchemite™ Focus Group meeting this month, bringing together a diverse set of chemicals, materials, and manufacturing organisations to discuss how to maximise value from Alchemite™ deep learning technology as they design and optimise products and processes.
This month saw the second invitation-only Focus Group session for project leaders currently using the Alchemite™ deep learning software or managers planning significant projects to apply the technology. The meeting enabled Alchemite™ user organisations to review the latest product developments and to provide input to Intellegens development plans, ensuring that these remain focused on the key needs of industry. It also further built up the Alchemite™ user community, sharing experiences across across varied applications. Attendees came from sectors including material production, aerospace, chemicals, formulated products, printing technology, and batteries. Thanks are especially due to the case study presenters from OCAS NV (optimising steel properties) and The AMRC (tooling for composites).
The group saw new features including: the ability to suggest initial experiments where no data is available; new ‘dimensionality reduction’ and ‘sensitivity’ plots, which help users to understand the machine learning models; and new controls on the relationships between data, which provide a way to introduce physical laws into a model. There was discussion of interesting work using Alchemite™ to gain insights from image data, and a number of upcoming new features were reviewed – look out for more information on these in future Intellegens newsletters.
The closing discussion focused on some of the challenges of implementation and roll-out, and how they can be overcome. Alchemite™ has now been successfully deployed on Azure, AWS, and Google Cloud infrastructures. The importance of establishing a clear business case for the use of machine learning in order to help drive projects was also covered, with the top business benefit among the group being identified as reducing the number of experiments required to complete projects.