How the Alchemite™ algorithm empowers machine learning solutions for R&D
This white paper provides a methodological and mathematical background for scientists interested in what underlies the unique capabilities of the Alchemite™ machine learning software.
Executive Summary
Machine learning is increasingly applied to accelerate R&D in sectors such as materials, chemicals, life sciences, and formulated products. Alchemite™, developed originally at the University of Cambridge and, since 2017, at Intellegens, is an algorithm tuned for the challenges found in these research areas. Notably, Alchemite™ works well with real experimental and process data, which is typically sparse and noisy, on which other machine learning methods fail. Another strength is in handling the uncertainty calculations that can be vital, for example, when prioritising experimental work. This white paper lifts the lid on some of these features, providing methodological and mathematical background for scientists interested in what underlies the unique capabilities of the Alchemite™ software.
