Abstract
Scientific workflows facilitate automation, reuse, and reproducibility of scientific data management and analysis tasks. Scientific
workflows are often modeled as dataflow networks, chaining together processing components (called actors) that query, transform, analyse, and visualize scientific datasets. Semantic annotations relate data and actor schemas withconceptual information from a shared ontology, to support scientific workflow design, discovery, reuse, and validation inthe presence of thousands of potentially useful actors and datasets. However, the creation of semantic annotations is complexand time-consuming. We present a calculus and two inference algorithms to automatically propagate semantic annotations through workflow actors described by relational queries. Given an input annotation α and a query q, forward propagation computes an output annotation α′; conversely, backward propagation infers α from q and α′.
Users
Please
log in to take part in the discussion (add own reviews or comments).