Exa4mind – Extreme Analytics for Mining Data spaces
Extreme Data with
the Best of Three Worlds:
Databases, Supercomputing & European Distributed Data
The EXA4MIND architecture has the goal of enabling data analytics workflows from data ingestion, over preprocessing to database management and evaluation by the user through massive queries. For Extreme Data mining, processing, and analytics, the data are made available on world-class supercomputers, and results are ingested back into our databases. Significant intermediate and final results are made available as per the scope of European FAIR Data Management Systems. Both control of the workflow control itself and the management of data flows will leverage advanced orchestration systems.
Using a co-design approach driven by all partners, the EXA4MIND Extreme Data platform will be gradually implemented and increasingly utilised by four Application Cases during the project.
Scientific
Application Case
Industry
Application Case
SME
Agricultural
Application Case
SME
Health
Application Case
Scientific
Application Case
Systematic Improvement of Molecular Simulations and Accuracy
What?
To develop ADAMS4SIMS, an innovative Automated Data Mining System for Systematic Improvement of Molecular Dynamics (MD) simulations accuracy and predictability, that will combine services and offline and online data sources.
What for?
- Creation of a massive storage and processing methodology for Extreme Data from molecular-dynamics simulations and respective observations to improve the development of force-field descriptions.
- Demonstration of the benefit of using massive amounts of FAIR data in molecular dynamics.
- User-friendly data mining interfaces for understanding force-field parameters relevant for simulation outcome.
- Generation of strongly-improved simulation results on a daily basis.
Industry
Application Case
Massively Automated Annotation & Evaluation of Automotive Camera Recordings
What?
To develop methods, AI models and frameworks, including technologies for Advanced Driver Assistance Systems (ADAS), Extreme Data annotation, search and analytics.
What for?
- Development of a self-supervised learning approach for annotating Lidar / front-facing camera data for ADAS, where AI models are trained to annotate the data to avoid tedious work.
- Implementation of a (meta-)database allowing smart queries based on weather, speed, location, etc.
- Analysis of root causes of ADAS failures, and automatic recognition of risky situations in this respect, with the help of the Extreme Data Database (EDD) and advanced analytics.
Led by Valeo
SME _ Agricultural
Application Case
Smart Farming/Viticulture with Sensors & Satellite Imagery
What?
To prove, emphasise and maximise the benefits of EXA4MIND platform for small-medium sized enterprises, focused on agriculture.
What for?
- Easy and quick ingestion of large amounts of data.
- Efficient and space-saving re-distribution of metadata, computed data or ingested data.
- Integration of EXA4MIND platform functionality with the TerraviewOS platform.
- Enabling EXA4MIND to carry out specific/pluggable data transformation operations.
- Easy large-scale satellite data inclusion (i.e.Copernicus and other ESA data sources) in TerraviewOS.
Led by Terraview
SME - Health
Application Case
Secure Data Mining in Health Data
What?
To prove, emphasise and maximise the benefits of EXA4MIND platform for small-medium sized enterprises, focused on the health sector.
What for?
- Demonstration of the viability of EDD for Extreme Data processing application cases with health and societal datasets.
- Coupling of the EDD with a graphical interface allowing ALTRNATIV and other SMEs to easily explore and process Big Data (ingest, transform and load such data at scale, including conversion etc.) with a reduced learning curve.
Led by Altrnativ