Rediscovering BCE Loss for Uniform Classification
This work addresses classification tasks by improving feature uniformity and performance in open-set scenarios like face recognition, though it is incremental as it builds on existing BCE loss methods.
The paper tackles the problem of uniform classification by proposing a unified threshold for all samples and deriving a BCE loss with a bias to learn this threshold, resulting in models that achieve higher uniform and sample-wise classification accuracy across six datasets and three feature extraction models.
This paper introduces the concept of uniform classification, which employs a unified threshold to classify all samples rather than adaptive threshold classifying each individual sample. We also propose the uniform classification accuracy as a metric to measure the model's performance in uniform classification. Furthermore, begin with a naive loss, we mathematically derive a loss function suitable for the uniform classification, which is the BCE function integrated with a unified bias. We demonstrate the unified threshold could be learned via the bias. The extensive experiments on six classification datasets and three feature extraction models show that, compared to the SoftMax loss, the models trained with the BCE loss not only exhibit higher uniform classification accuracy but also higher sample-wise classification accuracy. In addition, the learned bias from BCE loss is very close to the unified threshold used in the uniform classification. The features extracted by the models trained with BCE loss not only possess uniformity but also demonstrate better intra-class compactness and inter-class distinctiveness, yielding superior performance on open-set tasks such as face recognition.