Turning uncertainty into good decisions for chemicals, materials, and formulation R&D
There are in-built uncertainties in the data that we collect and the predictions that we make when developing chemicals, materials, or formulations. But that should not stop us from making good decisions using this information. Indeed, in some circumstances, we can even extract useful knowledge from what appears to be noise in the data. In this webinar, Intellegens CTO Dr Gareth Conduit delves into how we can gain useful insights with the right tools and approaches to uncertainty in machine learning studies.
We’ll see how this requires machine learning methods that can generate useful models even from sparse and noisy data, discuss requirements for effective uncertainty quantification, and hear about leading-edge research on extracting value from noise in the design of new concretes. The aim of this work is to save time and cost, and identify the most productive pathways for experimental programs and new product development.
First broadcast: 14 June, 2022
Speaker: Dr Gareth Conduit, Intellegens