JUNE 18–22, 2017

Presentation Details

Name: Challenges in Machine Learning for Complex Physical Systems
Time: Wednesday, June 21, 2017
03:45 pm - 04:15 pm
Room:   Panorama 3 – DEEP LEARNING DAY
Messe Frankfurt
Breaks:03:15 pm - 03:45 pm Coffee Break
Speaker:   Christoph Angerer, NVIDIA
In this session, we will discuss the challenges faced when applying machine learning to modelling complex physical systems. GPUs have proved to be an ideal computational platform for training and executing both deep and convolutional neural networks. With more and more GPUs are installed in HPC 
systems, scientific computing centres see growing machine learning workloads. Indeed, NVIDIA has invested in SaturnV, a large GPU-accelerated cluster, (#28 on the November 2016 Top500 list) to support  internal machine learning projects. After a brief introduction to deep learning on GPUs, we will address a selection of open questions physicists may face when using deep learning for their work. Research is making progress towards answering these questions but there remains plenty to be done in the field by the  deep learning and physics communities.