LeoNet: A Machine Learning Method for Binary Pulsar Classification
Abstract
Binary pulsars provide valuable laboratories for testing theories of gravity, but orbital Doppler shifts complicate their detection. Fourier-domain acceleration and jerk searches address this challenge via matched filtering, but at substantial computational cost. We present LeoNet, a convolutional neural network that uses ten learnable filters to extract features of signals affected by Doppler shifts. The resulting ten-channel feature map provides a compact, lower-dimensional alternative to an explicitly sampled acceleration-jerk response grid and is analysed by a convolutional classifier to identify candidate signals. For simulated observations lasting 500 s, LeoNet achieves a mean relative reduction in false negative rate of 55.5% across five sampling intervals compared with the evaluated PRESTO acceleration-search configuration. TensorRT-optimised LeoNet processes each 500 s observation in 3.44-4.37 ms in FP32 on an NVIDIA H100 PCIe GPU across eight sampling intervals, including preprocessing, inference, and postprocessing. At a sampling interval of 128 microseconds, its mean processing time is 3.54 ms, compared with 1.767 s for PRESTO FDAS on an AMD EPYC 9825 CPU with search-frequency limits of 96-1000 Hz, corresponding to an approximately 499-fold speedup in the measured processing time. These results suggest that LeoNet has the potential to improve detection performance, while its millisecond-scale processing time supports its use as a candidate-identification stage in real-time binary pulsar search pipelines.
Related
- Multi-Method Machine Learning Classification of Astronomical Sources with Rubin Data Previews: a Spectroscopy-Free Approach with DP1 and DP2
- Machine learning-based classification of type 1 active galactic nuclei
- Data representation matters: Optimising machine learning for X-ray source classification by leveraging spatio-spectral information
- CausticFlow: An Efficient Machine Learning Framework Combining Neural Differential Equations and Normalizing Flows for Binary Microlensing Parameter Inference
- CausticFlow: An Efficient Machine Learning Framework Combining Neural Differential Equations and Normalizing Flows for Binary Microlensing Parameter Inference
BibTeX
@misc{gong2026leonet,
title = {{LeoNet: A Machine Learning Method for Binary Pulsar Classification}},
author = {Zhaocheng Gong and Jack White and Zeyu Yang and Jayanta Roy and Karel Adamek and Wesley Armour},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.00908},
eprint = {2610.00908},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.00908},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - LeoNet: A Machine Learning Method for Binary Pulsar Classification
AU - Zhaocheng Gong
AU - Jack White
AU - Zeyu Yang
AU - Jayanta Roy
AU - Karel Adamek
AU - Wesley Armour
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.00908
UR - https://arxiv.org/abs/2610.00908
AB - Binary pulsars provide valuable laboratories for testing theories of gravity, but orbital Doppler shifts complicate their detection. Fourier-domain acceleration and jerk searches address this challenge via matched filtering, but at substantial computational cost. We present LeoNet, a convolutional neural network that uses ten learnable filters to extract features of signals affected by Doppler shifts. The resulting ten-channel feature map provides a compact, lower-dimensional alternative to an explicitly sampled acceleration-jerk response grid and is analysed by a convolutional classifier to identify candidate signals. For simulated observations lasting 500 s, LeoNet achieves a mean relative reduction in false negative rate of 55.5% across five sampling intervals compared with the evaluated PRESTO acceleration-search configuration. TensorRT-optimised LeoNet processes each 500 s observation in 3.44-4.37 ms in FP32 on an NVIDIA H100 PCIe GPU across eight sampling intervals, including preprocessing, inference, and postprocessing. At a sampling interval of 128 microseconds, its mean processing time is 3.54 ms, compared with 1.767 s for PRESTO FDAS on an AMD EPYC 9825 CPU with search-frequency limits of 96-1000 Hz, corresponding to an approximately 499-fold speedup in the measured processing time. These results suggest that LeoNet has the potential to improve detection performance, while its millisecond-scale processing time supports its use as a candidate-identification stage in real-time binary pulsar search pipelines.
ER -