Pistoia Alliance October Virtual Conference: MELLODDY Project

This on-demand recording is part of the Pistoia Alliance Collaborate to Innovate Virtual Conference Week, October 19-23, 2020.  For more information about related events, please visit our online calendar.

Summary

MELLODDY (Machine Learning Ledger Orchestration for Drug Discovery) is a groundbreaking collaboration that has the potential to accelerate drug discovery and improve patient outcomes. This 17 partner consortium brings together data owners, technology experts, and academia to develop a platform using federated learning to train machine learning models across multi-partner datasets while ensuring privacy-preservation of both the data and the models. Indeed, the decentralization of each pharma partner’s highly proprietary databases of annotated chemical libraries allows them, for the first time, to collaborate in their core competitive space.

As the project progresses, the MELLODDY platform will learn from an unprecedented volume of more than one billion drug development data points and hundreds of terabytes of image data, including annotations on biological effects of more than ten million small molecules. By design, the platform aims to boost the predictive performance and applicability domain of the models so to better predict the chances of success that a compound has in the later stages of drug discovery and development. The technology used in MELLODDY will hopefully inspire the same or similar technologies to be used in drug development, clinical practice and other sectors where the potential for ML-driven innovation has been constrained by fragmented data ownership, and complex and heterogenous data types.

Source: Broadband Commission Report on Digital and AI in Health (§4.2)

Speakers

Access the Recording

We will email you the recording.