An important property for deep neural networks is the ability to perform robust out-of-distribution detection on previously unseen data. This property is essential for safety purposes when deploying models for real world applications. Recent studies show that probabilistic generative models can perform poorly on this task, which is surprising given that they seek to estimate the likelihood of training data. To alleviate this issue, we propose the exponentially tilted Gaussian prior distribution for the Variational Autoencoder (VAE) which pulls points onto the surface of a hyper-sphere in latent space. This achieves state-of-the art results on the area under the curve-receiver operator characteristics metric using just the log-likelihood that the VAE naturally assigns. Because this prior is a simple modification of the traditional VAE prior, it is faster and easier to implement than competitive methods.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    The Exponentially Tilted Gaussian Prior for Variational Autoencoders


    Beteiligte:
    Floto, Griffin (Autor:in) / Kremer, Stefan (Autor:in) / Nica, Mihai (Autor:in)

    Erscheinungsdatum :

    2021


    Medientyp :

    Preprint


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Gaussian Process Prior Variational Autoencoders

    Casale, Francesco Paolo / Dalca, Adrian V / Saglietti, Luca et al. | ArXiv | 2018

    Freier Zugriff

    Variational Autoencoders

    Ghojogh, Benyamin / Crowley, Mark / Karray, Fakhri et al. | Springer Verlag | 2022


    Deep Tracking Portfolios Using Autoencoders and Variational Autoencoders

    Urrego, Daniel Aragón / Nieto, Oscar Eduardo Reyes / Quimbayo, Carlos Andrés Zapata | Springer Verlag | 2024


    Certifiably Robust Variational Autoencoders

    Barrett, Ben / Camuto, Alexander / Willetts, Matthew et al. | ArXiv | 2021

    Freier Zugriff

    Mixed-curvature Variational Autoencoders

    Skopek, Ondrej / Ganea, Octavian-Eugen / Bécigneul, Gary | ArXiv | 2019

    Freier Zugriff