Dingo-gw

Latest version: v0.5.11

Safety actively analyzes 638361 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 4 of 4

0.3.1

What's Changed
* Native time delay geocenter by max-dax in https://github.com/dingo-gw/dingo-devel/pull/126
* Modify test for FD waveform generation when fmax not a power of two by mpuerrer in https://github.com/dingo-gw/dingo-devel/pull/127
* Initial version of dingo_pipe by stephengreen in https://github.com/dingo-gw/dingo-devel/pull/129

**Full Changelog**: https://github.com/dingo-gw/dingo-devel/compare/v0.3.0...v0.3.1

Comments
* `dingo_pipe` will work when run with `local = True` but needs slightly modifications to the DAG files to be able to properly request GPU resources.

0.3.0

This updates the code to correspond to the recent paper on neural importance sampling, https://arxiv.org/abs/2210.05686.

Main changes

* Rename `SamplesDataset` class to `Result`. This class contains a set of samples (generated using a `Sampler` class to sample from a trained network), which it can perform various operations on. In particular, all functionality relating to importance sampling is contained here.
* Add sampling of synthetic phase to as a method in `Result`. This generates a phase sample given all the other parameters, using the GW likelihood. It takes into account higher multipole modes in the model by caching the waveform modes, and reconstructing the polarizations at different phases.
* Make various improvements to importance sampling. When using GNPE, importance sampling now takes place in the joint space $(\theta, \hat\theta)$, where $\hat\theta$ denotes the GNPE proxy parameters.
* Add documentation.

Main components still missing

* Improved inference interface including Asimov integration
* Ability to change prior, data conditioning, waveform model, etc., during importance sampling
* Detector calibration marginalization
* Code for generating synthetic noise PSD datasets
* Tutorials

**Full Changelog**: https://github.com/dingo-gw/dingo-devel/compare/v0.2.0...v0.3.0

0.2.0

This release contains code for the main Dingo functionality, including generating training sets, training, and inference. It is based on the paper https://arxiv.org/abs/2106.12594.

Code structure

* The code has been split into `core` and `gw` components, with the `core` component contain neural-network code, as well as generic sampling code not specific to gravitational waves (GWs). The `gw` component contains GW-specific code, including waveform and noise generation.
* There are command-line scripts (e.g., `dingo_generate_dataset`, `dingo_train`, `dingo_analyze_event`) to execute the main Dingo tasks. These typically take as arguments `.yaml` files containing the settings for the run. Example settings files are contained in the [examples](/examples) folder.

Additional components

Beyond the ideas contained in https://arxiv.org/abs/2106.12594, this code has

* Importance sampling. This uses the GW likelihood and prior to assign importance weights to Dingo samples. Since GNPE networks do not provide the log probability, it trains an additional unconditional flow to model the distribution of generated Dingo samples. Time and phase marginalisation are also implemented for the likelihood.
* Additional GNPE functionality, in particular factoring out the TaylorF2 phase to simplify the data representation. (So far, this does not seem to give a performance improvement.)

Main components still missing

* Documentation
* Tutorials
* Improved phase marginalisation for higher-mode importance sampling
* Integration with LVK software pipelines

Page 4 of 4

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.