MachinE Learning Ledger Orchestration for Drug DiscoverY

Deliverables

Other (7)

Public-consortium prototyping platform for public data (year 3)

Publicconsortium prototyping platform for public data

Final release of structure processing script

Final release of structure processing script, one version for each run (M7, 19, 29)

Prototype of based test-train fold splitting algorithm (year 3)

Prototype of based testtrain fold splitting algorithm ready for integration in WP4 ML software One for prototype for 2nd and 3rd federated run each

Prototype of based test-train fold splitting algorithm

Prototype of based test-train fold splitting algorithm ready for integration in WP4 ML software. One for prototype for 2nd and 3rd federated run each (M18, 29)

Final release of structure processing script (year 3)

Final release of structure processing script one version for each run

Locality-sensitive hashing based test-train fold splitting code deployed at the pharma partners

Final release of structure processing script (year 2)

Final release of structure processing script, one version for each run

Documents, reports (24)

Publications of key scientific findings from the Project

Publications of key scientific findings from the Project and collation of all publications and other communication modalities M2436

Comprehensive technical documentation (year 3)

Comprehensive technical documentation updated version each year

Documentation material of the version 3

Documentation material of the version 3 v3 of the platform and platform prerelease for audit by WP5

Data preparation manual (year 3)

Data preparation manual one version for each run

Report on aggregate performance metrics for top hits and information gain (year 3)

Report on aggregate performance metrics for top hits and information gain

Documentation material including articles and use case description (draft)

Documentation material including articles and use case description draft and final version M30 36

Risk and Data Management Plan (year 3)

Risk and Data Management Plan; updates

Documentation material of the version 1

Documentation material of the version 1 (v1) of the platform, and platform pre-release for audit by WP5

Report on federated learning schemes (year 3)

Report on federated learning schemes

Risk and Data Management Plan (year 2)

Risk and Data Management Plan; updates

Report on central infrastructure setup

Report on central infrastructure setup including central services (container registries, CI/CD pipelines, GPU acceleration, and partner cluster networking)

Risk Management Plan

Risk Management Plan; updates (M3, 12, 24)

Report on model accuracy and applicability domain metrics (year 3)

Report on model accuracy metrics first year and model accuracy and applicability domain metrics second and third year

Initial Project communication plan and toolset (year 2)

Initial Project communication plan and toolset; updates

Report on aggregated metrics for yearly run (year 3)

Report on aggregated metrics for yearly run

Data preparation manual

Data preparation manual one version for each run M6 18 29

Data Management Plan

Data Management Plan; updates (M3, 12, 24)

Data preparation manual (year 2)

Data preparation manual one version for each run

Report on computational strategies for meta-optimization and advanced hyperparameter tuning (year 3)

Report on computational strategies for metaoptimization and advanced hyperparameter tuning

Project Management Plan

Initial Project communication plan and toolset (year 3)

Initial Project communication plan and toolset updates

Description of transversal coordination process and agile method for IT development cycle

Documentation material including articles and use case description (final)

Documentation material including articles and use case description draft and final version

Documentation material of the version 2

Documentation material of the version 2 (v2) of the platform, and platform pre-release for audit by WP5

Websites, patent fillings, videos etc. (1)

Initial Project communication plan and toolset

Initial Project communication plan and toolset; updates (M12, 24, 36)

Publications

Conference proceedings (1)

Privacy-Preserving and Bandwidth-Efficient Federated Learning: An Application to In-Hospital Mortality Prediction

Author(s): Raouf Kerkouche, Gergely Acs, Claude Castelluccia, Pierre Geneves
Published in: ACM Conference on Health, Inference, and Learning (CHIL), 2021, Page(s) 25-35, ISBN 978-1-4503-3307-8
Publisher: ACM
DOI: 10.1145/3450439.3451859

Peer reviewed articles (3)

Don’t Overweight Weights: Evaluation of Weighting Strategies for Multi-Task Bioactivity Classification Models

Author(s): Lina Humbeck, Tobias Morawietz, Noe Sturm, Adam Zalewski, Simon Harnqvist, Wouter Heyndrickx, Matthew Holme, Bernd Beck
Published in: MDPI molecules, 2021, 26 (22), 2021, ISSN 1420-3049
Publisher: Multidisciplinary Digital Publishing Institute (MDPI)
DOI: 10.3390/molecules26226959

Splitting chemical structure data sets for federated privacy-preserving machine learning

Author(s): Simm, Jaak; Humbeck, Lina; Zalewski, Adam; Sturm, Noe; Heyndrickx, Wouter; Moreau, Yves; Beck, Bernd; Schuffenhauer, Ansgar
Published in: Journal of Cheminformatics, 2021, vol 13, 2021, ISSN 2573-2293
Publisher: Springer
DOI: 10.1186/s13321-021-00576-2

ChemGrapher: Optical Graph Recognition of Chemical Compounds by Deep Learning

Author(s): Martijn Oldenhof, Adam Arany, Yves Moreau, Jaak Simm
Published in: Journal of Chemical Information and Modeling, 60/10, 2020, Page(s) 4506-4517, ISSN 1549-9596
Publisher: American Chemical Society
DOI: 10.1021/acs.jcim.0c00459