Deep Learning for Multimessenger Astrophysics: Real-time Discovery at Scale
Agenda
All activities will be held at the NCSA Building, 1205 W. Clark St., Urbana, Illinois, unless otherwise noted.
Tuesday, October 16
7:00 pm
Wednesday, October 17
8:00 am
Registration
9:00 am
Welcome
Bill Gropp, Eliu Huerta
9:15 am
Discussion: Signal-processing Algorithms to Enable Real-time Gravitational Wave Discovery / Status Update and Outstanding Challenges
Chair: Bernard Schutz
Participants:
- Maya Fishback (KICP, UChicago)
- Leo Singer (NASA Goddard)
- Zsuzsa Marka (Columbia)
- Erik Katsavounidis (MIT)
10:45 am
Break
11:00 am
Discussion: Signal-processing Algorithms to Enable Real-time Electromagnetic and Astro-particle Followups of Gravitational Wave Candidates / Status Update and Outstanding Challenges
Chair: Federica Bianco
Participants:
- Philip Cowperthwaite (Harvard)
- Rahul Biswas (UW)
- Yue Shen (University of Illinois)
- Igor Andreoni (Caltech)
- Alexander Olivas (UMD)
- Francisco Forster (CMM)
12:30 pm
Lunch
1:45 pm
Discussion: Modeling and Simulations of MMA Sources / Status Update and Outstanding Challenges
Chair: Gabrielle Allen
Participants:
- Jim Annis (FNAL)
- Katrin Heitmann (ANL)
- Andrea Derdzinski (Columbia)
- Stu Shapiro (University of Illinois)
3:15 pm
Break
3:30 pm
Symposium: Emergent Trends to Simulate MMA Sources
Chair: Roland Haas
Participants:
- Daniel Siegel (Columbia)
- Jonah Miller (LANL)
- Manuela Campanelli (RIT)
- Zach Etienne (WVU)
5:00 pm
Parallel Breakouts / Preliminary Report for MMA Sources: Emergent Trends for Detection and Modeling
6:30 pm
Conference dinner sponsored by NVIDIA
Thursday, October 18
8:00 am
Demos
Participants: NCSA Gravity Group, Blue Waters SEAS Team, NVIDIA
NCSA scientists describe their work at the interface of AI & HPC for gravitational wave astrophysics, and demonstrate its potential to enable real-time gravitational wave discovery.
9:00 am
Break
9:15 am
Symposium and Discussion: Emergent Trends in Machine and Deep Learning for Signal Processing Algorithms that Are Relevant for MMA
Chair: Elise Jennings
Participants:
- Joshua Bloom (UC Berkeley)
- Matthew Graham (Caltech/ZTF)
- Daniel George (NCSA)
- Alex Schwing (University of Illinois)
10:45 am
Break
11:00 am
Discussion: Reproducibility, Portability and Interpretability of Deep Learning Algorithms / Heuristic vs Rigorous Approaches in Deep Learning
Chair: Jane Zhao
Participants:
- Elise Jennings (ANL)
- Jinjun Xiong (IBM)
- Prasanna Balaprakash (ANL)
- Ashish Mahabal (Caltech)
12:30 pm
Lunch
1:45 pm
Symposium and Discussion: Status Update of Existing and Future NSF and DoE Facilities Relevant for Deep Learning at Scale
Chair: Minsik Cho
Participants:
- Bill Kramer (NCSA)
- Jack Wells (ORNL)
- Tom Gibbs (NVIDIA)
- Vlad Kindratenko (NCSA)
- Tim Williams (ANL)
- John Towns (NCSA)
3:15 pm
Break
3:30 pm
Discussion: Open Data, Software, and Workflows / Policies and Good Practices Sharing Tools and Data in Real-time to Enable Scientific Discovery in the MMA Era
Chair: Daniel S. Katz
Participants:
- Bruce Berriman (Caltech)
- Matt Turk (University of Illinois)
- Aaron Saxton (NCSA)
- Erik Katsavounidis (MIT)
5:00 pm
Break
5:15 pm
Preparation for Final Reports: Deep Learning at Scale for MMA / Development of Software for Community Efforts
Friday, October 19
8:00 am
Symposium: Status of Current Efforts to Adopt and Deploy MMA Algorithms and Brokers / Target and Observation Managers
Chair: Eliu Huerta
Participants:
- Gautham Narayan (STScI)
- Hongyu Shen (ECE and NCSA, University of Illinois)
- Francisco Forster (Millennium Institute for Astrophysics)
- Leo Singer (NASA)
- Etienne Bachelet (LCO)
8:50 am
Break
9:00 am
Discussion: Roadmap to Adopt and Deploy Machine Learning Algorithms for MMA Discovery Campaigns
Chair: Federica Bianco
Participants:
- Gautham Narayan (STScI)
- Hongyu Shen (ECE and NCSA, University of Illinois)
- Francisco Forster (Millennium Institute for Astrophysics)
- Leo Singer (NASA)
- Etienne Bachelet (LCO)
9:50 am
Break
10:00 am
Discussion: Cyberinfrastructure Experts Share Insights about Best Practices to Develop Deep Learning Algorithms and Share Them with the Scientific Community / Importance of Credit and Citation
Chair: Daniel S. Katz
Participants:
- Barnabas Poczos (CMU)
- Bruce Berriman (Caltech)
- Kenton McHenry (NCSA)
- Jonah Miller (LANL)
- Kyle Chard (UChicago)
- Brigitta Sipocz (University of Washington)
Noon
Lunch
1:00 pm
Discussion: Roadmap of Community Efforts to Develop Mature Software Stacks to Simulate MMA Sources
Chair: Charles Gammie
Participants:
- Jonathan Zrake (Columbia)
- Zach Etienne (WVU)
- Roland Haas (NCSA)
- Jane Zhao (ECE and NCSA, University of Illinois)
- Anshu Dubey (ANL)
2:20 pm
Break
2:30 pm
Discussion: Design Roadmap for Future Collaborations
Chairs: Gabrielle Allen, Eliu Huerta, Daniel S. Katz
Participants:
- Alexander Olivas (UMD)
- Bernard Schutz (AEI/Cardiff)
- Ed Seidel (University of Illinois)
- John Towns (NCSA, University of Illinois)
- Joaquin Vieira (Astronomy, University of Illinois)
- Jinjun Xiong (IBM)
- Mark Neubauer (Physics, University of Illinois)
4:00 pm
End of Workshop
Shuttle takes participants to Willard Airport (Expected arrival: 4:30 pm, which is plenty of time to catch a 5:30 flight)
Hackathon
Sponsored by NVIDIA
Logistics
- Start: October 17 at 9:00 am
- End: October 18 at 12:00 am
Prizes will be given October 19 at 3:30 pm.
This workshop is funded by the NSF through award NSF 1848815 and
co-sponsored by NVIDIA and the Argonne Leadership Computing Facility.