A Unified Latent Space Disentanglement VAE Framework with Robust Disentanglement Effectiveness Evaluation

Xiaoan Lang, Md Mostafizer Rahman, Fang Liu

Evaluating and interpreting latent representations, such as variational autoencoders (VAEs), remains a significant challenge for diverse data types, especially when ground-truth ge…