ONLINE -- MSA Seminar: Terabyte-scale image analysis with HPC-enabled Deep Learning for building a map of the human brain

Modular Supercomputing Architectures Seminar

Speaker: Christian Schiffer, INM-1, Forschungszentrum Jülich

This talk will introduce our work on automated classification of human brain areas in high-resolution microscopic scans of histological brain sections, with a particular focus on the technical workflow running on the HPC systems at JSC.

It will further cover basics of common machine learning frameworks in Python, which enable easy use of NVIDIA GPUs for Deep Learning and other machine learning tasks.
Finally, the talk will cover some more advanced techniques for Deep Learning with large amounts of data, such as distributed training, mixed precision training and gradient checkpointing.

This seminar is offered online via BigBlueButton at https://webconf.fz-juelich.de/b/and-2jh-qj4 .

In case of difficulties with BigBlueButton, the seminar will be switched to DFNconf streaming at https://www.conf.dfn.de/stream/nr5hm44jym4c4