Evaluating Disentanglement in Generative Models Without Knowledge of Latent Factors

C Holtz, G Mishne, A Cloninger - … , Algebraic and Geometric …, 2022 - proceedings.mlr.press
Probabilistic generative models provide a flexible and systematic framework for learning the
underlying geometry of data. However, model selection in this setting is challenging …

Tc-vae: Uncovering out-of-distribution data generative factors

C Meo, A Goyal, J Dauwels - arxiv preprint arxiv:2304.04103, 2023 - arxiv.org
Uncovering data generative factors is the ultimate goal of disentanglement learning.
Although many works proposed disentangling generative models able to uncover the …

Measuring hallucination in disentangled representations

H Benazha, S Ayache, H Kadri… - 2024 International Joint …, 2024 - ieeexplore.ieee.org
Disentanglement is a key challenge in representation learning as it may enable several
downstream tasks including edition operation at a high semantic level or privacy-preserving …

Gradient Adjusting Networks for Domain Inversion

E Sheffi, M Rotman, L Wolf - Scandinavian Conference on Image Analysis, 2023 - Springer
StyleGAN2 was demonstrated to be a powerful image generation engine that supports
semantic editing. However, in order to manipulate a real-world image, one first needs to be …