Mixture separability loss in a deep convolutional network for image classification
This addresses the early saturation issue in image classification networks, offering a potential improvement for researchers and practitioners, though it appears incremental as it modifies the loss function rather than introducing a new paradigm.
The paper tackles the problem of early saturation in deep convolutional networks for image classification by proposing a novel cost function called mixture separability loss (MSL), which updates weights even when most training images are correctly predicted, and experiments show it deepens learning and yields promising results on datasets like SVHN, CIFAR, and ICVL.
In machine learning, the cost function is crucial because it measures how good or bad a system is. In image classification, well-known networks only consider modifying the network structures and applying cross-entropy loss at the end of the network. However, using only cross-entropy loss causes a network to stop updating weights when all training images are correctly classified. This is the problem of the early saturation. This paper proposes a novel cost function, called mixture separability loss (MSL), which updates the weights of the network even when most of the training images are accurately predicted. MSL consists of between-class and within-class loss. Between-class loss maximizes the differences between inter-class images, whereas within-class loss minimizes the similarities between intra-class images. We designed the proposed loss function to attach to different convolutional layers in the network in order to utilize intermediate feature maps. Experiments show that a network with MSL deepens the learning process and obtains promising results with some public datasets, such as Street View House Number (SVHN), Canadian Institute for Advanced Research (CIFAR), and our self-collected Inha Computer Vision Lab (ICVL) gender dataset.