RenSho – Creating a comprehensive platform to enhance reproducibility in data-science driven research with Renku and Shogun
- Gunnar Rätsch, ETH Zürich
The first goal of this project is to provide Renku with a scheme for integrating external ML libraries, to allow broadening of the methodology that is accessible through the platform. This includes providing schemes on how to connect external ML libraries to Renku’s internal workflow tools, and user templates that empower non-experts to embedding third party methodology into their workflows without detailed technical knowledge as well as defining an ontology to be able to formally process results of machine learning models within Renku. The second goal of this project is to specifically integrate Shogun within Renku. This will provide the significant efforts that went into reproducibility and transparency of the various models/algorithms within Shogun to the Renku community.