As promised, another public SCIMMA telecon tomorrow (Tuesday 15th September, 3pm Eastern, 2pm Central, 1pm Mountain, noon Pacific). We will hear from another NSF HDR-funded MMA project with presentations from Eliu Huerta (NCSA, UIUC) and Dylan Sheldon Rankin (MIT)
Title: Advancing Science with Accelerated Machine Learning
Abstract: We present recent advances of our interdisciplinary and multi-institutional project to harness machine learning, extreme scale computing, and scientific visualization to address computational grand challenges in big-data physics experiments. We summarize outcomes of our AI educational program at the interface of academia and industry, discuss our vision towards the creation of a Harnessing the Data Revolution Institute and present the first realization of both an FPGA and GPU based Machine Learning heterogeneous computing workflow for real-time processing of LHC collisions. We comment on the extensibility and feasability of this design and present new results on ML-based neutrino reconstruction.
See you tomorrow!
Adam Brazier for SCIMMA