Integrated approaches to deep learning in gravitational-wave signal modeling and statistical inference

The detection and characterization of gravitational-wave signals in data from contemporary and next-generation detectors can be computationally challenging or even prohibitive, due to the accuracy requirements imposed on certain waveform models and the high-dimensional posterior sampling required for Bayesian inference. I discuss several promising strategies for streamlining the process with an integrated model-to-inference framework, making use of techniques from the deep-learning paradigm.