Rocco Sedona's Successful PhD Defense: Making HPC for Deep Learning More Accessible in Earth Observation

On May 4, 2023, Rocco Sedona defended his PhD thesis, "Scalable Deep Learning for Remote Sensing with High Performance Computing," at the Faculty of Industrial Engineering, Mechanical Engineering and Computer Science at the University of Iceland. As his supervisor, I can confidently say that his work has made a significant difference in the interdisciplinary of field of Machine Learning with HPC for Earth Observation (EO).

Rocco's work aims to solve a big problem in EO - the amount of time it takes to use advanced Deep Learning (DL) models because of the large volume of data needed for their training. His solution is to use distributed deep learning on supercomputers to speed up the training process. This approach makes it faster to develop these models, making them more effective and easier to use.

He has also made improvements in land cover (LC) mapping. By using data from various remote sensing (RS) missions, he has improved the speed at which areas can be studied. This approach not only makes LC classification better but also provides useful information about changes in the landscape. This information can help scientists and policymakers make important decisions about things like climate change and resource planning.

Rocco's work makes it easier for the RS community to use DL methods with HPC for EO applications, which will help push forward research and development in this interdisciplinary area.

During his defense, Rocco showed a deep understanding of both the scientific and computational parts of his research. He answered difficult questions clearly and confidently, showing his skills as a researcher and innovator.

As his supervisor along with Prof. Morris Riedel and Prof Matthias Book, I'm proud of Rocco's hard work and the contribution he has made, and I'm excited to see how his work will influence the future of the field.

‍

More information (PhD thesis, recording of the defence)

‍