NERSCPowering Scientific Discovery Since 1974

Big Data Summit 2019: AI and HPC Convergence for Science

The Big Data Summit 2019, was hosted by NERSC on July 9, 2019.

All presentations and recordings of the event are available for download from a Google folder or through links in the agenda below.

The event featured results from advanced Data Analytics and Machine Learning projects on the NERSC Cori system. Ongoing projects being led by NERSC, Intel, Cray, and five Intel® Parallel Computing Centers (represented by the University of Liverpool, University of California Berkeley, University of California Davis, New York University, and University of British Columbia) were featured in the talks. Experts from NERSC, Intel, and invited speakers covered Data Analytics and Data Management at Scale.

Agenda

8:30–9:00 a.m. Registration and Refreshments
9:00–9:45 a.m. Welcome and Agenda Review; NERSC Update – Data Analytics and Machine Learning (Prabhat, NERSC); Presentation: slides (PDF, PPTX), recording (audio-only, video)
9:45–10:15 a.m. Data-parallel Training of Generative Adversarial Networks on HPC Systems for HEP Simulations (Sofia Vallecorsa, CERN openlab); Presentation: slides (PDFPPTX), recording (audio-onlyvideo)
10:15–10:30 a.m. Break
10:30–11:00 a.m. Progress of HPC and AI Convergence (Victor Lee, Intel Corporation); Presentation: slides (PDF), recording (audio-onlyvideo)
11:00–11:30 a.m. AI, Numerical Models, and the Power of Approximation for Science (Mike Ringenburg, Cray); Presentation: slides (PDF), recording (audio-onlyvideo)
11:30 a.m.–12:30 p.m. Lunch
12:30–1:00 p.m. End-to-End Hierarchical Clustering with Graph Neural Networks (Joan Bruna and Nicholas Choma, New York University)Presentation: slides (PDF, PPTX), recording (audio-onlyvideo)
1:00–1:30 p.m. Deep Models and Probabilistic Inference for Astronomy (Jeffrey Regier and Runjing Liu, University of California – Berkeley)Presentation: slides (PDF), recording (audio-onlyvideo)
1:30–2:00 p.m. Towards Topological and Machine Learning Pattern Detection Methods in Spatiotemporal Climate Data (Vitaliy Kurlin and Grzegorz Muszynski, University of Liverpool); Presentation: slides (PDF, PPTX), recording (audio-onlyvideo)
2:00–2:30 p.m. Probabilistic Programming & Inference Compilation for Particle Physics (Gunes Baydin, Oxford University); Presentation: slides (PDF), recording (audio-onlyvideo)
2:30–3:00 p.m. Project DisCo: Physics-based Discovery of Coherent Structures in Spatiotemporal Systems (Adam Rupe, University of California – Davis); Presentation: slides (PDF), recording (audio-onlyvideo)
3:00–3:15 p.m. Break
3:15–3:35 p.m. Jupyter @ NERSC (Rollin Thomas, NERSC); Presentation: slides (PDF, PPTX), recording (audio-onlyvideo)
3:35–3:55 p.m. Enabling and Enhancing DL Workflows @ NERSC (Steve Farrell, NERSC); Presentation: slides (PDF), recording (audio-onlyvideo)
3:55–4:15 p.m. I/O for Deep Learning at Scale (Quincey Koziol, NERSC); Presentation: slides (PDF, PPTX), recording (audio-onlyvideo)
4:15–4:45 p.m. NERSC Machine Learning Survey Results (Mustafa Mustafa, NERSC); Presentation: slides (PDF), recording (audio-onlyvideo)
4:45–5:15 p.m. Community Feedback Session (Victor Lee, Intel Corporation); Discussion (audio-only, video)
5:15–5:30 p.m. Closing Remarks (Prabhat, NERSC); Recording (audio-only)