The paper investigates how autoencoder parameters reflect the statistical properties of their training data. By analyzing the spectral characteristics of the parameter matrices, it shows that singular values correspond to eigenvalues of the data covariance matrix, linking data and parameter spaces. Experiments on CIFAR‑10 and FashionMNIST demonstrate that these spectral vectors can accurately differentiate models trained on different data subsets without complex generation methods or access to the original samples.
arXiv:2608.29867v1 Announce Type: new
Abstract: Autoencoders are widely used for nonlinear dimensionality reduction and manifold learning. While most common implementations rely on both nonlinear enc...
By Louen Pottier, Louis Lesueur, Anders Thorin
arXiv:2602. 10680v2 Announce Type: replace-cross Abstract: Many real-world datasets contain hidden structure that cannot be detected by simple linear correlations between input features.
By Vicente Conde Mendes, Lorenzo Bardone, C\'edric Koller, Jorge Medina Moreira, Vittorio Erba, Emanuele Troiani, Lenka Zdeborov\'a
arXiv:2606. 25900v1 Announce Type: new Abstract: Variational Autoencoders (VAEs) belong to a family of autoencoders with probabilistic properties, making them well suited for generating data by producing a smooth and continuous latent space.
By Gananath R
Variational Autoencoders (VAEs) belong to a family of autoencoders with probabilistic properties, making them well suited for generating data by producing a smooth and continuous latent space. Despite being introduced over a decade ago, the method continues to be widely adopted in both research and industry for diverse applications.
arXiv:2607. 01275v1 Announce Type: cross Abstract: Variational Autoencoders (VAEs) commonly assume a standard isotropic Gaussian prior over the latent space, an assumption that often fails to capture the true distribution of latent representations for complex datasets.
By Qijun Chen, Shaofan Li