All publications

Unlocking New Paths for Efficient Analysis of Gravitational Waves from Extreme-Mass-Ratio Inspirals with Machine Learning

Bo Liang, Hong Guo, Tianyu Zhao, He Wang, et al.

All 14 authors

Bo Liang1, Hong Guo2, Tianyu Zhao, He Wang, Herik Evangelinelis, Yuxiang Xu, Chang Liu, Manjia Liang, Xiaotong Wei, Yong Yuan, Minghui Du11, Peng Xu12, Weiliang Qian, Ziren Luo

  • 1 Equal contribution
  • 2 Equal contribution
  • 11 Corresponding author
  • 12 Corresponding author

JournalChinese Physics Letters 42(8), 081101 (2025)

First public: Published online:

  • Parameter inference
  • AI for science

Research summary

Extreme-mass-ratio inspirals have complex waveforms and difficult posterior structure. This study applies flow matching and continuous normalizing flows to accelerate Bayesian parameter estimation. Simulated analyses compare learned posteriors with conventional sampling, discussing both improved exploration of broad parameter spaces and the need to refine some learned uncertainty estimates.

Extended notes

The method targets the computational difficulty of EMRI posterior estimation. The published study notes that learned distributions can require refinement with conventional Bayesian sampling.

Sources