Adapting to noise distribution shifts in flow-based gravitational-wave inference
This enables low-latency analyses of gravitational waves by allowing deep learning techniques to adapt to evolving detector characteristics, though it is incremental as it builds on existing DINGO networks.
The paper tackles the problem of adapting deep learning models for gravitational-wave parameter estimation to changing noise distributions, by developing a probabilistic model to forecast future noise power spectral densities (PSDs). Using data from observing runs O2 and O3, they trained a DINGO network that performed accurate inference on 37 real events throughout O3.
Deep learning techniques for gravitational-wave parameter estimation have emerged as a fast alternative to standard samplers $\unicode{x2013}$ producing results of comparable accuracy. These approaches (e.g., DINGO) enable amortized inference by training a normalizing flow to represent the Bayesian posterior conditional on observed data. By conditioning also on the noise power spectral density (PSD) they can even account for changing detector characteristics. However, training such networks requires knowing in advance the distribution of PSDs expected to be observed, and therefore can only take place once all data to be analyzed have been gathered. Here, we develop a probabilistic model to forecast future PSDs, greatly increasing the temporal scope of DINGO networks. Using PSDs from the second LIGO-Virgo observing run (O2) $\unicode{x2013}$ plus just a single PSD from the beginning of the third (O3) $\unicode{x2013}$ we show that we can train a DINGO network to perform accurate inference throughout O3 (on 37 real events). We therefore expect this approach to be a key component to enable the use of deep learning techniques for low-latency analyses of gravitational waves.