Navigation und Service

IAS-Seminar "Towards scalable AI: Opportunities & challenges of distributed Deep Learning"

26.01.2021 10:00 Uhr
26.01.2021 11:00 Uhr
online via Zoom


Dr. Peter Labus, Large Scale Deep Learning Team Lead, Fraunhofer Institute for Industrial Mathematics ITWM, Kaiserslautern, Germany


Much of the recent progress in Deep Learning has been achieved through datasets and Deep Neural Networks (DNNs) growing in size, thereby reaching the limits of the available hardware and software capabilities.

In this talk, I will give an overview of the opportunities High Performance Computing can provide to scale and improve Deep Learning methods. I will particularly focus on our recent work using Neural Architecture Search to minimize the energy consumption for DNN inference on FPGAs, as well as our Deep Learning framework Tarantella for distributed DNN training on CPUs and GPUs.


Dienstag, 26. Januar 2021, 10.00-11.00 Uhr


Online via Zoom

Meeting ID: 996 0616 7252
Passcode: 2NJVdi1k