labrador: A domain-optimized machine-learning tool for gravitational wave inference
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910118412550144 |
|---|---|
| author | Roulet, Javier Crisostomi, Marco Thomas, Lucy M. Chatziioannou, Katerina |
| author_facet | Roulet, Javier Crisostomi, Marco Thomas, Lucy M. Chatziioannou, Katerina |
| contents | Fast and reliable inference of gravitational-wave source parameters is crucial for analyzing large catalogs that are reaching the size of hundreds of detections, and for identifying short-lived electromagnetic counterparts. Neural posterior estimation has emerged as a powerful inference method, where the model is trained on simulated gravitational-wave data at considerable computational cost, but thereafter enables extremely fast and inexpensive inference at test time. Here, we extend this approach by incorporating domain-specific physical insights and methods in the model architecture. These include compressing the data by heterodyning against a reference waveform chosen via approximate likelihood maximization, removing parameter degeneracies through tailored coordinate systems, and eliminating known multimodalities by folding the parameter space. As a result, the network is approximately equivariant to changes in the source parameters, and achieves a reduced training cost and improved model interpretability. Our implementation, called labrador, can be trained end-to-end on a 1-day timescale on $\sim 10^2$ CPU cores and a V100 GPU, achieving a median importance-sampling efficiency of 1% on quadrupolar, aligned-spin signals in a broad mass range (chirp mass $\mathcal M \in 1\text{-}50\,\mathrm{M}_\odot$, mass ratio $q > 0.1$). labrador is the first neural inference code to achieve extensive coverage of long-duration signals with secondary masses $m_2 < 10\,\mathrm{M}_\odot$, rendered possible by its equivariance property. Among our novel contributions is a numerically stable procedure that enables neural posterior estimation when the simulation and inference priors differ. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_08897 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | labrador: A domain-optimized machine-learning tool for gravitational wave inference Roulet, Javier Crisostomi, Marco Thomas, Lucy M. Chatziioannou, Katerina General Relativity and Quantum Cosmology High Energy Astrophysical Phenomena Instrumentation and Methods for Astrophysics Fast and reliable inference of gravitational-wave source parameters is crucial for analyzing large catalogs that are reaching the size of hundreds of detections, and for identifying short-lived electromagnetic counterparts. Neural posterior estimation has emerged as a powerful inference method, where the model is trained on simulated gravitational-wave data at considerable computational cost, but thereafter enables extremely fast and inexpensive inference at test time. Here, we extend this approach by incorporating domain-specific physical insights and methods in the model architecture. These include compressing the data by heterodyning against a reference waveform chosen via approximate likelihood maximization, removing parameter degeneracies through tailored coordinate systems, and eliminating known multimodalities by folding the parameter space. As a result, the network is approximately equivariant to changes in the source parameters, and achieves a reduced training cost and improved model interpretability. Our implementation, called labrador, can be trained end-to-end on a 1-day timescale on $\sim 10^2$ CPU cores and a V100 GPU, achieving a median importance-sampling efficiency of 1% on quadrupolar, aligned-spin signals in a broad mass range (chirp mass $\mathcal M \in 1\text{-}50\,\mathrm{M}_\odot$, mass ratio $q > 0.1$). labrador is the first neural inference code to achieve extensive coverage of long-duration signals with secondary masses $m_2 < 10\,\mathrm{M}_\odot$, rendered possible by its equivariance property. Among our novel contributions is a numerically stable procedure that enables neural posterior estimation when the simulation and inference priors differ. |
| title | labrador: A domain-optimized machine-learning tool for gravitational wave inference |
| topic | General Relativity and Quantum Cosmology High Energy Astrophysical Phenomena Instrumentation and Methods for Astrophysics |
| url | https://arxiv.org/abs/2604.08897 |