The use of deep neural networks in real-world applications require well-calibrated networks with confidence scores that accurately reflect the actual probability. However, it has been found that these networks often provide over-confident predictions, which leads to poor calibration. Recent efforts have sought to address this issue by focal loss to reduce over-confidence, but this approach can also lead to under-confident predictions. While different variants of focal loss have been explored, it is difficult to find a balance between over-confidence and under-confidence. In our work, we propose a new loss function by focusing on dual logits. Our method not only considers the ground truth logit, but also take into account the highest logit ranked after the ground truth logit. By maximizing the gap between these two logits, our proposed dual focal loss can achieve a better balance between over-confidence and under-confidence. We provide theoretical evidence to support our approach and demonstrate its effectiveness through evaluations on multiple models and datasets, where it achieves state-of-the-art performance. Code is available at https://github.com/Linwei94/DualFocalLoss


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Dual Focal Loss for Calibration


    Beteiligte:
    Tao, Linwei (Autor:in) / Dong, Minjing (Autor:in) / Xu, Chang (Autor:in)

    Erscheinungsdatum :

    2023


    Anmerkungen:

    ICML 2023 Accept


    Medientyp :

    Preprint


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    On Focal Length Calibration from Two Views

    Sturm, P. / IEEE | British Library Conference Proceedings | 2001



    COMPOSITE FOCAL PLANE ARRAY (CFPA) IMAGING SYSTEM GEOMETRIC CALIBRATION

    ANTONIADES YIANNIS / EDWARDS JONATHAN / CHESTER DAVID | Europäisches Patentamt | 2024

    Freier Zugriff

    Automated Focal Loss for Image based Object Detection

    Weber, Michael / Furst, Michael / Zollner, J. Marius | IEEE | 2020


    AUTOMATED FOCAL LOSS FOR IMAGE BASED OBJECT DETECTION

    Weber, Michael / Fürst, Michael / Zöllner, J. Marius | British Library Conference Proceedings | 2020