Joel Strickland from Intellegens will be presenting at the American Institute of Chemical Engineers Spring Meeting. Joel will be discussing the practicalities of applying AI methods to aid innovation in chemical and process research.
If you’re attending the conference, why not contact us to arrange a meeting?
Abstract
Collaborative Intelligence: A Federated Framework for Accelerated Product Innovation and Secure Data Sharing
Joel Strickland, Intellegens, Cambridge, United Kingdom
To address increasing demands for cost-efficient, high-performance products, this framework enables companies to combine proprietary models without compromising data confidentiality. Using a shared calibration dataset, companies can leverage a mix of unique and overlapping inputs and property measurements to derive actionable insights across varied conditions. Independently trained models handle missing data through imputation and calibrate predictive accuracy by aligning uncertainty, enabling resilient predictions even with sparse inputs.
This adaptive approach prioritizes prediction reliability, enhancing experimental planning and reducing unnecessary physical testing and equipment use. By drawing on each model’s strengths to optimize specific design spaces, it creates a fast feedback loop that continuously improves product predictions as data is collectively refined. This secure, data-driven method has significant potential in fields such as chemicals, pharmaceuticals, and food, offering a path to enhanced predictive accuracy and accelerated product development while safeguarding proprietary information.
Session: Analytics & AI – Industry Perspective
