WP7a early use case requirements and story boards
LarKC platform from WP7a perspective
The section presents our understanding about the LarKC platform and the different plugins and their usage. Some of the plugin types may be improperly used (like retrieval vs selection), so I urge to correct me.
Early LLD and demonstration plan
- Load the LLD data (e.g. as result of offline retrieval) - Different data sources are transformed to RDF.
- Perform reasoning for data integration (abstraction) - Perform light-weight (tractable) reasoning to remove redundant data (i.e. OWL-Horst)
- Extract semantic annotations and store document-related metadata (retrieval) - Use text-mining algorithms for named-entity recognition.
- Extract relations from text (retrieval of new knowledge) - Apply text-mining relation extraction algorithms to generate new entities.
- Perform reasoning to detect inconsistent extractions (reasoning) - Improve the efficiency of the information extraction algorithms by applying DL-like reasoning to detect inconsistencies.
- Evaluate queries against the documents, involving knowledge from LLD (reasoning/selection) - Retrieve all resources (textual documents or database entries) to relates to lung diseases (COPD is specific type of lung diseases) and cancer (in all its specializations)
Some other important requirements to be considered:
Efficient communication between the different plugins should be possible (extreme volumes of information would be transfered)
Security will not be considered as an issue in WP7a use case.
Storyboards
The current version of the story boards is still not public. An internal ?AstraZeneca audit of the document has to certify that D7a.1.1 deliverable does not disclose sensitive company information. All current storyboards could be realized using the described work flow. If new requirements arise the page will be updated immediately.