Inetum contributes to the BD4QoL project from two different perspectives, from a technical one, with the design and provision of a secure and interoperable data lake equipped with cutting-edge big data technologies and operations, enabling the deployment of analytical models to support patients QoL improvement, and from an open perspective, analyzing different strategies to allow the use of BD4QoL by the greatest number of survivors, seeking the viability and sustainability of the provision of BD4QoL services.
The project will make use of an infrastructure that supports the required features for the cross-functional teams involved in delivering the machine learning workflows, from the ingestion and preparation of the data, through the exploratory analysis of the datasets and the experimentation environments to discover the value of the features, the computation of derived features to help the convergence of the algorithms and finally the publication of models and the monitoring of their performance.
The infrastructure will provide different warehouses with specific characteristics for each stage of information processing, with a very high capacity staging area suitable for the landing of raw data, data stores prepared for the EDA (exploratory data analysis) workbench and high-performance in-memory data structures for real-time data processing.
All the information handled in the project will be published in the form of a data hub that will feature different access APIs with appropriate levels of protection, through the definition of specific access realms for the different entities.
In order to enable the construction of machine learning models, high performance devices such as GPUs that will be provisioned on-demand when analytical workloads request it. This availability will be elastic, so that it can grow or shrink on demand and thus optimise the computing costs inherent to this type of project. The infrastructure will also allow to package, deploy and serve the trained models in an industrialisation model that automates their monitoring and fine-tuning, as well as model output publishing to consumer applications and services.
Inetum has started the BD4QoL strategic market study, analyzing the different value propositions for the different users and ecosystem providers involved in the throat and neck cancer problem.
From an open perspective, several strategies are analyzed to allow the use of BD4QoL by the greatest number of survivors, seeking the viability and sustainability of the provision of BD4QoL services to improve their living conditions after cancer through the use of technology and medical assistance monitored in real time.
