The Universal COVID-19 Pipeline was presented at the NVIDIA GTC 21 conference.
The prestigious event NVIDIA GTC’21 (GPU Technology Conference), started on April 12thl with the NVIDIA CEO’s keynote and ran continuously over five days across all time zones. It was entirely virtual and free to attend online even if registration was needed. There were 1400+ talks and more than 150000 registered attendees.
Prof. Marco Aldinucci of the University of Turin was invited to present the Universal COVID-19 Pipeline developed by the Confederation of Laboratories for Artificial Intelligence Research in Europe (CLAIRE) task force on AI & COVID-19 during the first COVID-19 outbreak. The group, composed of fifteen researchers in complementary disciplines (Radiomics, AI, and HPC) and led by Prof. Aldinucci, investigated the diagnosis of COVID-19 pneumonia assisted by Artificial Intelligence (AI) and supported by enabling technologies such as High Performance Computing (HPC) and Clouds. More information on the COVID-19 initiative are available in the Springer Communication “The CLAIRE COVID-19 initiative: approach, experiences and recommendations” published on February 2021.
The talk presented approaches and results of addressing the diagnosis of COVID-19 pneumonia using AI techniques and other enabling technologies like HPC and Cloud. In practice, the Universal Pipeline is a way to describe a broad class of analysis pipelines in a formalized way. The pipeline was designed and executed using StreamFlow, the novel Workflow Management System developed in the DeepHealth project and explicitly designed to support reproducibility, a major issue in computational research, and portability across different platforms. The talk ended providing a broader perspective on the lesson learned and future steps on the convergence between AI and HPC.