Healthcare applications: a spot for GPU, FPGA, CSD, and In-Memory computing accelerators

Abstract:
In the last decade, research in the healthcare scenario exposed the intersection of many disciplines like medicine, bioinformatics, electronics engineering, statistics, and so on. The explosion of Machine- and Deep-Learning methods for medical data analysis (e.g., imaging data, patient records, etc.) raised concerns about the development of computationally yet energy efficient High-Performance Computing (HPC) pipelines to surpass the state-of-the-art platforms based on CPUs only. This a preferential spot for the development of accelerators based on heterogeneous technologies such as GPU, FPGA, Computational Storage Drives (CSD), and eventually In-Memory Computing architectures. This presentation will show the benefits of integrating those technologies in at-scale HPC environments devoted to healthcare applications such as COVID-19 detection, CT image segmentation for aortic calcification detection, and patient survival analysis enrolled in a retrospective study on heart-related diseases.

Speaker bio:
Cristian Zambelli received the M.Sc. and Ph.D. in Electronics Engineering and Engineering Science from the University of Ferrara, Ferrara, Italy, in 2008 and 2012, respectively. In 2015, he joined the Department of Engineering at the same institution where he currently holds an Associate Professor position in Electronics. His current research interests include electrical characterization and reliability modeling at an array level of different non-volatile memory technologies such as Flash (Planar and 3D) for memory reliability/performance trade-off exploitation in Solid State Drives (SSDs) and emerging concepts such as Resistive RAM (RRAM) oriented towards In-memory and neuromorphic computing accelerators. He also works in developing acceleration solutions based on GPU, FPGA, and Computational Storage Devices for High-Performance Computing applications in different scenarios such as multi-physics applications, precision medicine, and omics.