Explaining the ‘black box’: deep learning in drug discovery

Recent years have seen a remarkable rise in the number and scope of artificial intelligence and machine learning (especially deep learning) algorithms for small molecule discovery. But how can these methods provide new and transformative insights into our discovery data? How can we uncover complex relationships within our data, such as mechanisms of action or adverse outcome pathways, and guide compound design using explainable AI?
In this webinar, Dr Daniel Barr, Senior Application Scientist at Optibrium, explores, using case studies, the workings and impact of our deep learning imputation platform, Cerella™.
Through practical, real-world examples, Daniel demonstrates:
1) How deep learning imputation outperforms traditional QSAR methods when predicting activities and properties without sacrificing interpretability.
2) How these models capture more than twice the information about the relationships between endpoints than correlation analysis.
3) How Cerella’s approach improves prediction accuracy by excluding extraneous data and utilising indirect correlations that other methods miss.