-
Notifications
You must be signed in to change notification settings - Fork 11
Program
NOTE: SCHEDULE AND TOPICS SUBJECT TO CHANGE
Day's goal: Understand how to do a parameter estimation analysis and how to interpret results.
9:00 - 9:10: Welcome [Collin Capano & Ian Harry]
9:10 - 9:50: Intro to parameter estimation in gravitational wave astronomy [Matthew Pitkin]
- Basics of Bayesian Inference
- Application to gravitational-wave astronomy
9:50 - 10:30: Gravitational waves from compact binary coalescence [Sebastian Khan]
- Physics, parameters, and degeneracies
10:30 - 11:00: Break
11:00 - 11:45: Markov-chain Monte Carlo (MCMC) [Vivien Raymond]
- Basic theory
- Ensemble sampling
- Parallel tempering
- Thermodynamic integration
- Implementations
- Practical considerations: convergence tests, thinning
11:45 - 12:30: Nested sampling [John Veitch]
- Basic theory
- Implementations
- Practical considerations: choosing number of live points, tolerance; ...
12:30 - 14:00: Lunch
14:00 - 15:00: Introduction to PyCBC [Alex Nitz] [Tutorial]
- Tutorial covering basics of waveform generation, matched filtering
- Data conditioning
- PSD estimation
- Practical considerations: wrap around, inverse spectrum truncation, etc.
15:00 - 15:15: Break
15:15 - 16:15: Introduction to PyCBC Inference [Collin Capano]
- Code overview
- How to set up a simple test run
- How to set up a CBC analysis
- Output file format
- Plotting results
16:15 - 16:30: Break
16:30 - 17:15: Defining priors and parameters [Collin Capano]
- Variable vs sampling vs waveform parameters
- The distributions module
- Using transforms
17:15 - 18:00: Using the workflows [Chris Biwer]
- Standard workflow
- Injection (percentile-percentile test) workflow
Day's goal: Understand how the code works and how to modify it for new projects.
9:00 - 9:45: MCMC implementations and settings [Collin Capano]
- Emcee/EmceePT
- Epsie
9:45 - 10:30: Nested sampling implementations and settings [Sumit Kumar]
- CPNest
- PyMultiNest
10:30 - 11:00: Break
11:00 - 11:45: A primer on python classes [Collin Capano] [Tutorial]
- What is a class
- Class inheritance
- Abstract base classes
11:45 - 12:30: Code road map [Collin Capano] [Tutorial]
- What's called where
- Module structure
12:30 - 14:00: Lunch
14:00 - 15:30: Available models and their use [Alex Nitz] [Tutorial]
- Basic API
- Available models
- How to add a new model
15:30 - 16:00: Break
16:00 - 16:45: LIGO/Virgo Collaboration development: Bilby [Vivien Raymond]
16:45 - 17:15: Parallelization [Alex Nitz]
- Multiprocessing
- MPI
17:15 - 17:30: How to contribute using github [Ian Harry]
Day's goal: Establish improvements to work on for future.
9:00 - 9:45: Evidence calculation with parallel tempering: pitfalls, tips, and improvements [Steven Reyes]
9:45 - 10:00: Break
10:00 - 10:45: Web-based tools for workflow visualization [Chris Biwer]
10:45 - 11:00: Break
11:00 - 12:00: PESummary [Charlie Hoy]
12:00 - 13:30: Lunch
The afternoon will mostly be unstructured project and discussion time, giving people a chance to ask questions about anything they found unclear from the first two days. Multiple parallel discussions may occur. Possible discusions points:
- What can be improved?
- How to implement on a cluster? Continue to use Condor? Slurm? Other options?
The ICG is committed to fostering an environment of dignity and respect. We ask that all speakers and participants read the ICG's code of conduct to help ensure that the workshop is friendly and welcoming to all participants.