Date: 16th October, 2025
Time: 4pm BST | 11am EST | 8am PT | 5pm CEST
If you’re using predictive models in your molecule design and optimisation, an accurate uncertainty estimate can be just as important as the result itself. Where prediction uncertainties are low, you can trust your results and prioritise compounds with confidence, saving valuable time and resources in redundant synthesis and testing. But what’s often missed is that exploring areas where the uncertainties are high can reveal even better compounds.
In this webinar, we’ll explore practical methods to quantify model uncertainties and demonstrate how good uncertainty estimates enable you to extract value even from seemingly poor models.
Using case studies and examples, you’ll see how uncertainty-aware modelling helps you make better decisions and avoid costly mistakes, and we’ll show how including uncertainties in multi-parameter optimisation can enable you to find compounds with a better chance of success. In the real world, no model is perfect. The key is in extracting as much value as possible from what you have.
