On Focal Loss for Class-Posterior Probability Estimation: A Theoretical Perspective
| Autor: | Nuttapong Chairatanakul, Masashi Sugiyama, Nontawat Charoenphakdee, Jayakorn Vongkulbhisal |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
FOS: Computer and information sciences
Computer Science - Machine Learning Contextual image classification Calibration (statistics) business.industry Computer science Posterior probability Machine Learning (stat.ML) Pattern recognition Object detection Machine Learning (cs.LG) Transformation (function) Statistics - Machine Learning Classifier (linguistics) Pattern recognition (psychology) Minification Artificial intelligence business |
| Zdroj: | CVPR |
| DOI: | 10.1109/cvpr46437.2021.00516 |
| Popis: | The focal loss has demonstrated its effectiveness in many real-world applications such as object detection and image classification, but its theoretical understanding has been limited so far. In this paper, we first prove that the focal loss is classification-calibrated, i.e., its minimizer surely yields the Bayes-optimal classifier and thus the use of the focal loss in classification can be theoretically justified. However, we also prove a negative fact that the focal loss is not strictly proper, i.e., the confidence score of the classifier obtained by focal loss minimization does not match the true class-posterior probability and thus it is not reliable as a class-posterior probability estimator. To mitigate this problem, we next prove that a particular closed-form transformation of the confidence score allows us to recover the true class-posterior probability. Through experiments on benchmark datasets, we demonstrate that our proposed transformation significantly improves the accuracy of class-posterior probability estimation. Comment: 57 pages |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|