On the Shape of Latent Variables in a Denoising VAE-MoG: A Posterior Sampling-Based Study
This highlights the importance of posterior-based validation for generative models, particularly in domain-specific applications like gravitational wave analysis, though it is incremental in focusing on validation rather than new methods.
The study tackled the problem of evaluating latent space reliability in a denoising VAE-MoG trained on gravitational wave data, finding that despite accurate signal reconstruction, there was a clear mismatch in latent representations when comparing encoder outputs to posterior samples.
In this work, we explore the latent space of a denoising variational autoencoder with a mixture-of-Gaussians prior (VAE-MoG), trained on gravitational wave data from event GW150914. To evaluate how well the model captures the underlying structure, we use Hamiltonian Monte Carlo (HMC) to draw posterior samples conditioned on clean inputs, and compare them to the encoder's outputs from noisy data. Although the model reconstructs signals accurately, statistical comparisons reveal a clear mismatch in the latent space. This shows that strong denoising performance doesn't necessarily mean the latent representations are reliable highlighting the importance of using posterior-based validation when evaluating generative models.