Jianan Liu
- PhD Student, Affiliate
- University of Auckland
I am a PhD candidate in Statistics at the University of Auckland, and my research tackles a long-standing bottleneck in gravitational-wave astronomy: producing accurate, low-latency noise models. I develop stochastic-gradient variational Bayesian algorithms that estimate detector power-spectral densities (PSDs) in seconds, enabling noise models to be refreshed continuously during LIGO and Virgo observing runs. Building on this capability, I am constructing a joint inference framework that fits both the PSD and the source parameters of compact-binary signals in a single pass, eliminating the bias that arises when noise is treated as fixed. I am rigorously evaluating my approach against the traditional Welch estimator on real LVK data and adapting it for the wider bandwidth and lower noise floor anticipated for the Einstein Telescope.
View external profileView all profiles