Transform the models used to hunt the Higgs boson to develop everyday applications, from car insurance fraud to early diagnosis.
This is one of the objectives of the researchers working in Spoke 2 of the Icsc, the National Center for Research in High Performance Computing, Big data and Quantum Computing, a section dedicated to 'Fundamental research and Space economy'.
"I have been working on one of the LHC experiments for some time and I am currently working on trying to understand how to optimize the time and computational resources required by current experiments, such as Atlas, and by future experiments, which will be located at the future large circular accelerator Fccee, whose implementation is being evaluated by Cern", said Adelina D'Onofrio, of the Naples section of the National Institute of Nuclear Physics.
These are innovative IT solutions for processing enormous amounts of data that require a lot of time and energy and for this reason it becomes essential to find possible 'shortcuts' capable of making these processes more sustainable.
"Thus, solutions created for the world of basic research can also find useful applications in other fields," said D'Onofrio.
One of the major tools used to reduce processing loads is the so-called parallelization, i.e. the possibility of finding new ways to divide the operations to be performed into blocks and process them in a parallel way so as to reduce time and consumption.
Obviously machine learning algorithms also have an important role, an example being the 'anomaly detection' methods capable of identifying some anomalies in enormous volumes of data that can be fundamental.
"Methods - added the researcher - which can for example be used in agriculture, on satellite images to recognize important data for individual crops or to analyze user databases to identify potential car insurance fraud".
Indeed, this is precisely one of the objectives of the center created thanks to 320 million euros allocated by the Pnrr to create a distributed supercomputing infrastructure available to both researchers and companies.
Reproduction reserved © Copyright ANSA