Timezone: »

 
A Federated Learning benchmark for Drug-Target Interaction
Filip Svoboda · Gianluca Mittone · Nicholas Lane · Pietro Lió

@
in Machine Learning in Structural Biology Workshop »

Aggregating pharmaceutical data in the drug-target interaction (DTI) domain has the potential to deliver life-saving breakthroughs. It is, however, notoriously difficult due to regulatory constraints and commercial interests. This work proposes the application of federated learning, which we argue to be reconcilable with the industry's constraints, as it does not require sharing of any information that would reveal the entities' data or any other high-level summary of it. When used on a representative GraphDTA model and the KIBA dataset it achieves up to 15% improved performance relative to the best available non-privacy preserving alternative. Our extensive battery of experiments shows that, unlike in other domains, the non-IID data distribution in the DTI datasets does not deteriorate FL performance. Additionally, we identify a material trade-off between the benefits of adding new data, and the cost of adding more clients.

Author Information

Filip Svoboda (University of Cambridge)
Gianluca Mittone (University of Turin)
Gianluca Mittone

Gianluca Mittone received his Bachelor’s Degree in Computer Science in 2017 with a thesis on the handling of exceptions in Description Logics, proposing the implementation of an algorithm for the automatic revision of ontologies exploiting a Typicality operator. He also received the Master’s Degree in Computer Science in 2019 with a master thesis on a novel distributed approach for deep learning, named NNT (Nearest Neighbours Training), which takes advantage of a locally synchronous approach to achieve a better trade-off between computational time and learning results. After eight months as a research engineer at the Computer Science Department of the University of Turin, he is now a Ph.D. student in Modeling and Data Science at the same university, working on different projects involving HPC and Machine Learning techniques.

Nicholas Lane (University of Cambridge and Samsung AI)
Pietro Lió (University of Cambridge)

More from the Same Authors