In this workshop, we will provide tutorials on how to use PyCBC Inference to do parameter estimation with gravitational waves. Tutorials will involve going through jupyter notebooks that highlight how the various methods and functions in PyCBC Inference work, and how to use the exectuables to do gravitational-wave parameter estimation. We will also review basics of Bayesian inference and stochastic sampling, and their application to gravitational waves.

We will hold the meeting via Zoom. We will email participants the connection information.

To run the tutorials, we will be using SciServer. If you don't already have one, please create a SciServer account (it's free) before the meeting.

Note: we will go through the following instructions on Thursday morning, before going through the first tutorial. Feel free to do this ahead of time if you like, however. Go to https://apps.sciserver.org/compute/.

1. Click "Create container". Give it a name. In the "Compute Image" drop-down menu make sure "SciServer Essentials" is selected. Then click "Create."
2. Click on the container you just created; this will open a new tab in your browser that is a Jupyter notebook interface.
3. All of the tutorials we will use are in the PyCBC-Tutorials repository. Clone the PyCBC-Tutorials repository into your SciServer container: Click "New" -> "Terminal". This will open another tab that with a bash terminal in it. Change directory into "workspace" by typing cd workspace. Now type:

   git clone https://github.com/gwastro/PyCBC-Tutorials.git
   

   This will download a copy of this repository to your directory on SciServer.
4. We will be using the Workshop2020 tag of PyCBC-Tutorials. To ensure that you are using that tag, after the clone, change directory into "PyCBC-Tutorials" by typing cd PyCBC-Tutorials. Now type: git checkout tags/Workshop2020 -b Workshop2020 This will checkout a local branch of the Workshop2020 release.
5. Go back to your tab with the Jupyter notebook. Click on the "PyCBC-Tutorials" directory. From there you can navigate to the tutorial you want to view (using the itinerary below as a guide). When you click on a tutorial (files that end in .ipynb), a new tab will open with the tutorial open in it. All of the code needed is installed at the beginning of the tutorials.

All times are CEST.

• 11:00 - 12:00: Lecture
  • Introduction to Bayesian Inference and PyCBC Inference keynote version, pdf version
• 12:00 - 12:15: Setting up Sciserver
• 12:15 - 14:30: break
• 14:30 - 17:30: Tutorials
  • Learn about models and do parameter estimation in a notebook (Inference #1)
  • PyCBC Inference code road map (Inference #2)
  • Using PyCBC Inference on an analytic test (Inference #3)

• 11:00 - 12:00: Tutorial
  • How to estimate the parameters of a binary black hole merger (Inference #4)
• 12:00 - 13:00: break
• 13:00 - 15:00
  • How to load, plot, and manipulate results (Inference #5)
  • Advanced configuration settings (Inference #6)
  • How to add custom waveforms to PyCBC (Inference #7)
• 15:30 - 16:25: How to use PyCBC Inference in Condor workflow Note: this session will focus on how to use the workflow on the Atlas cluster for members of the AEI Hannover, but will be instructive for anyone that has access to a condor-enabled cluster.
  • Analyzing multiple events at once
  • Doing percentile-percentile tests and/or injection studies
• 16:25 - 16:30: Closing Remarks