Architecture Acceleration of Machine Learning Gravitational Waveform Surrogate ModelsNew

This research paper targets gravitational-wave physics, specifically waveform modeling for compact binary coalescences. It presents differentiable and hardware-accelerated implementations of the mlgw and mlgw-bns machine-learning surrogate models in the JAX ecosystem. The framework supports JIT compilation, vector parallelization, native GPU acceleration, and automatic differentiation while retaining the original surrogate training infrastructure.