Pseudo Labels for Single Positive Multi-Label Learning
This work addresses the high annotation cost problem in multi-label image classification for researchers and practitioners, offering a cost-effective solution that is incremental in improving label efficiency.
The paper tackles the challenge of single positive multi-label learning, where only one positive label per image is available, by introducing Pseudo Multi-Labels, a method that uses a teacher network to generate pseudo-labels for training a student network, achieving performance close to models trained on fully-labeled data.
The cost of data annotation is a substantial impediment for multi-label image classification: in every image, every category must be labeled as present or absent. Single positive multi-label (SPML) learning is a cost-effective solution, where models are trained on a single positive label per image. Thus, SPML is a more challenging domain, since it requires dealing with missing labels. In this work, we propose a method to turn single positive data into fully-labeled data: Pseudo Multi-Labels. Basically, a teacher network is trained on single positive labels. Then, we treat the teacher model's predictions on the training data as ground-truth labels to train a student network on fully-labeled images. With this simple approach, we show that the performance achieved by the student model approaches that of a model trained on the actual fully-labeled images.