Thursday 30 April 2026, by Corentin Louis
The EXTRACT project (<https://extract-project.eu/> ) aims to design a distributed data-mining software platform that can efficiently and smoothly orchestrate tasks across a continuum of computing resources, including edge devices, cloud systems, and high-performance computing centres. One use-case of EXTRACT is TASKA (Transient Astrophysics with an SKA Pathfinder), in which this platform will be used to process massive astronomy datasets, particularly those generated by the NenuFAR instrument, a SKA Pathfinder developed in Nançay, France. NenuFAR has the potential to produce up to 109 PB of raw data annually, necessitating on-site processing using carefully designed pipelines to reduce the dataset to approximately 1 PB per year.
The TASKA project is structured into four distinct use cases:
(a) Real-time detection of high-resolution events within beamformer data streams, seamlessly integrated with artificial intelligence algorithms.
(b) The development of a modular and adaptable radio imaging pipeline, capable of operating on local, cloud, or high-performance computing resources.
(c) The design of a novel dynamic radio imager, employing video reconstruction techniques and artificial intelligence, as an integral component of the radio imaging pipeline.
(d) The creation of a spatial-temporal-spectral feature-matching tool, facilitating the integration of detected events with images generated by the imaging pipeline.
The current status and results of TASKA activities demonstrate their application to NenuFAR data processing and their relevance for the SKA Regional Centre network.