Hackathon Plot Plan for FastEMRIWaveforms - znasipak/FastEMRIWaveforms-Soton-Hackathon-2025 GitHub Wiki

Hackathon Plot Plan for FastEMRIWaveforms

This document outlines the key plots for the paper associated with the new release of the FastEMRIWaveforms package. The plots are categorized based on their focus: accuracy, speed, and science with EMRIs. Each section can be expanded with additional plots as necessary. Each plot should have an accompanying text describing what the plot wants to communicate. The goal is to implement two Python scripts per plot: one to generate the data to make the plot and one for generating the plot and add as well a markdown.

Paper plot repo can be found here.

1. Accuracy

These plots will assess how closely the FastEMRIWaveforms results match reference models.

Trajectory Module

  • Plot Name: Phase Drift Comparison

    • X-axis: Time (years)
    • Y-axis: Phase error (radians)
    • Goal: Compare phase evolution of FEW-generated trajectories against a post-Newtonian (PN) benchmark after several years of inspiral. Understand the transition from PN to relativistic fluxes.

  • Plot Name: _Trajectory Mismatch vs. ODE error Lorenzo _

    • X-axis: ODE error
    • Y-axis: Dephasing against high-resolution reference trajectory
    • Goal: Show how decreasing ODE error improves trajectory accuracy but with diminishing returns.

  • Plot Name: Interpolated Fluxes vs. Numerical fluxes_

    • X-axis: p,e
    • Y-axis: relative difference
    • Goal: Which errors are we making in the interpolation

Amplitude Module

  • Plot Name: Mode Amplitude Errors
    • X-axis: Mode index (e.g., ( n ), ( m ), ( \ell ))
    • Y-axis: Relative error in amplitude compared to Teukolsky in sparse points
    • Goal: Highlight discrepancies.

Waveform Module

  • Plot Name: Waveform Mismatch vs. Eccentricity/Spin
    • X-axis: Eccentricity/Spin
    • Y-axis: Mismatch with reference waveforms
    • Goal: Show the accuracy of the waveform module at different eccentricities.

2. Speed

These plots will measure the computational efficiency of different modules.

Trajectory Module

  • Plot Name: ODE Integration Time vs. ODE error/Accuracy
    • X-axis: ODE error
    • Y-axis: Computation time (ms)
    • Goal: Demonstrate the trade-off between accuracy and speed in trajectory generation. Justification for the default.

Amplitude Module

  • Plot Name: Interpolation Time per Mode
    • X-axis: Number of modes
    • Y-axis: Generation time (ms)
    • Goal: Compare the computational cost of different interpolation methods for mode amplitudes.

Waveform Module

  • Plot Name: Waveform Generation Time vs. Mode Content
    • X-axis: Fraction of retained modes (( \epsilon ))
    • Y-axis: Time to generate a waveform (ms)
    • Goal: Show the impact of mode selection on waveform generation speed.

3. Science (Physical Insights & Performance in Astrophysical Contexts)

These plots will demonstrate the impact of the new waveform model on the science of asymmetric binaries.

Trajectory Module

  • Plot Name: Inspiral Lifetime vs. Spin/Eccentricity
    • X-axis: Black hole spin and eccentricity ( a, e )
    • Y-axis: Inspiral duration (years)
    • Goal: Show how the spin of the primary black hole and eccentricity affects the inspiral time, relevant for LISA observation windows and formation scenarios, how does this compare to previous Peters and Matthew or 5PN?

Amplitude Module

  • Plot Name: SNR Comparison Between AAK and Kerr Models

    • X-axis: Primary mass ( M ) (in ( M_\odot ))
    • Y-axis: Accumulated SNR
    • Goal: Show where simpler models overestimate or underestimate the signal strength compared to fully relativistic Kerr waveforms.

  • Plot Name: SNR per Harmonic Kerr Model

    • X-axis: Harmonic mode n
    • Y-axis: Harmonic mode m
    • Goal: Color coded pixels in n,m to show the importance of different harmonic contributions for the Kerr model.

Waveform Module

  • Plot Name: Sky Localization Error vs. Eccentricity/Duration

    • X-axis: Eccentricity/Duration
    • Y-axis: Uncertainty in sky localization (deg²)
    • Goal: Show how including more waveform modes improves LISA’s ability to localize sources.

  • Plot Name: High eccentric SNRs

    • X-axis: Eccentricity
    • Y-axis: SNR, lifetime
    • Goal: Do these sources matter?

  • Plot Name: Redshift Horizon Plot

    • X-axis: Primary mass ( M ) (in ( M_\odot ))
    • Y-axis: Maximum redshift ( z )
    • Goal: Show how far LISA can detect EMRIs in redshift space.

  • Plot Name: MCMC Parameter Estimation: AAK vs. Kerr

    • X-axis: Parameter values (e.g., mass, spin, eccentricity)
    • Y-axis: Posterior distributions
    • Goal: Compare the parameter estimation accuracy of AAK vs. Kerr waveforms.

  • Plot Name: Fisher Matrix: Minimum Measurable Eccentricity and Spin

    • X-axis: Signal-to-noise ratio (SNR)
    • Y-axis: Minimum detectable eccentricity/spin
    • Goal: What is the smallest eccentricity and spins accurately.

Goals of the day:

  • Niels: I want the trajectory to stop the trajectory on certain surface.
  • Christian: developer guide for the trajectory.
  • Susanna: Extending the trajectory 1PA and beyond.
  • Patrick: Working on the tutorials, provide feedback! Also happy to do other tasks
  • Chris: Working on the tutorials, provide feedback! Also happy to do other tasks
  • Ollie: trash french GPUs
  • Loic: tutorials + understand the structure of FEW, hybrid
  • Eccentric, tutorials + understand the structure of FEW, hybrid, fluxes to provide
  • Full inspiral merger ringdown model
  • Jonathan: Plots, Waveform integration
  • Phil: resonance, generic Kerr interpolation, relation with other variables (convention doc)
  • Christian: GPUs for trajectory
  • Zach: Plots accuracy, build the 4D spline, what is the initial generic grid. Needed stuff:
  • Plots Assign yourself one single task you want to get done by today, realistically.