|Task 7.1: Coordination and links to other WPs|
|Task 7.2: Innovations in platform, sensor and data interoperability|
|Task 7.3: Innovative ecosystem sensors and sensor package developments|
|Task 7.4: Enabling data science from innovative observation services and systems|
|Task 7.5: JERICO e-Infrastructure|
|Task 7.6: In situ demonstration of sensor packages|
Enabling data science from innovative observation services and systems
M1-M36 (Lead: CNR, partners: PLOCAN, IFREMER, UPC, NORCE, CNRS, Hereon, SYKE, SOCIB):
This task defines and develops data science methodologies for the interpretation and modelling of relevant biological and ecological processes based on the data produced by the innovative sensors experimented within JERCO-S3. The proposed methodologies will be designed to be included in the JERICO e-infrastructure (Task 7.5).
Subtask 7.4.1 State of the art (M1 – M6): As a fast evolving field, the state of the art on data science methodologies will be reviewed for marine data processing, modelling and interpretation. Special interest will be on application contexts related to ecosystem coastal monitoring.
Subtask 7.4.2 Data science methodologies definition and implementation (M7 – M20): novel methodologies based on evolutionary computing and deep learning, for knowledge discovery, data processing, modelling and interpretation, will be defined and developed mainly focusing on : flow cytometry analysis; image recognition and classification of macro and mega fauna; multivariate analysis and modeling of time series combining environmental, physical, biological, biogeochemical and metabarcoding data. Ground truth datasets will be defined for the training and validation of machine learning algorithms and for the general assessment of the proposed data processing approaches.
Subtask 7.4.3: Intelligent services integrated into JIIM (M21 – M36): the data science methodologies implemented will be used to define and implement the intelligent services for the JIIM. Such intelligent services will be capable to trigger JIIM sensors and sampler, and adapt sensor configurations (e.g. data sampling frequency) for the automated adaptation of the JIIM behavior to the dynamic relevant environmental conditions.
Subtask 7.4.4: Data science methodologies and Intelligent services integrated into the e-infrastructure (M21 – M33): the JERICO community will integrate the defined data science methodologies developed in task 7.4.2 into the VA e-infrastructure for easy access. This will enable deployment at European scale, foster implementation feedback and sustainability beyond the project time frame.
Task 7.4: D7.8: Intelligent services and data science methodologies for the JIIM and the VA e-infrastructure.