We benchmark the robustness of maximum likelihood based uncertainty estimation methods to outliers in training data for regression tasks. Outliers or noisy labels in training data results in degraded performances as well as incorrect estimation of uncertainty. We propose the use of a heavy-tailed distribution (Laplace distribution) to improve the robustness to outliers. This property is evaluated using standard regression benchmarks and on a high-dimensional regression task of monocular depth estimation, both containing outliers. In particular, heavy-tailed distribution based maximum likelihood provides better uncertainty estimates, better separation in uncertainty for out-of-distribution data, as well as better detection of adversarial attacks in the presence of outliers.


    Access

    Download


    Export, share and cite



    Title :

    Maximum Likelihood Uncertainty Estimation: Robustness to Outliers


    Contributors:

    Conference:



    Publication date :

    2022-03-01


    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English




    Uncertainty Quantification Using Maximum Likelihood: Experimental Validation

    Fonseca, Jose / Friswell, Michael / Mottershead, John et al. | AIAA | 2005


    Uncertainty identification by the maximum likelihood method

    Fonseca, José R. | Online Contents | 2005


    MAXIMUM-LIKELIHOOD GPS PARAMETER ESTIMATION

    Progri, Ilir F. | Online Contents | 2005


    Maximum-Likelihood Parameter-Estimation Algorithm

    Eldred, D. B. / Hamidi, M. / Rodriguez, G. | NTRS | 1986


    Maximum likelihood estimation for orbit determination

    BARRETT, M. / BRIERLEY, S. | AIAA | 1986