Miðbiksmat í reikniverkfræði - Liang Tian

Zoom hlekkur: https://eu01web.zoom.us/j/64011126091

Heiti ritgerðar: N/A (Advancing Large-Scale Remote Sensing Applications with Scalable AI on Modular Supercomputers)

Nemandi: Liang Tian

Doktorsnefnd:
Dr. Morris Riedel, prófessor við Iðnaðarverkfræði-, vélaverkfræði- og tölvunarfræðideild HÍ
Dr. Gabriele Cavallaro, dósent við Rafmagns- og tölvuverkfræðideild HÍ
Dr. Rocco Sedona, aðstoðarstjóri Hermunar- og gagnaverkfræðistofu (SDL) gervigreindar og vélræns náms fyrir fjarskynjun, Forschungszentrum Jülich, Þýskalandi

Ágrip á ensku

Remote Sensing (RS) is a process of sensing (detecting and monitoring) the physical characteristics of an area by means of measuring the reflected and emitted radiation from a satellite or an airborne sensor. It helps to classify different physical features that occupy the surface of the Earth (e.g., land-cover classes) and to describe the use of the land surface by humans (i.e., land-use classes). The use of Machine Learning (ML) and Deep Learning (DL) in the context of classification has been the topic of research for quite some time now due to the possibility of generating accurate classification results. The large-scale data generated from RS technologies requires significant computing power to process and analyze. To address these needs, parallel algorithms that can scale on heterogeneous and high-performance computing technologies, including High-Performance Computing (HPC) platforms, will be used. HPC refers to using supercomputers or computer clusters to perform complex computational tasks.