Semi-supervised Learning with a Teacher-student Network for Generalized Attribute Prediction
This work addresses the challenge of ambiguous attribute hierarchies in vision applications, offering a semi-supervised solution for tasks like fashion attribute prediction, but it appears incremental as it builds on existing multi-task and distillation techniques.
The paper tackles the problem of visual attribute prediction in computer vision, which suffers from class imbalance and label sparsity, by proposing a multi-teacher-single-student (MTSS) approach that achieves competitive performance on fashion attribute benchmarks and improves robustness and cross-domain adaptability.
This paper presents a study on semi-supervised learning to solve the visual attribute prediction problem. In many applications of vision algorithms, the precise recognition of visual attributes of objects is important but still challenging. This is because defining a class hierarchy of attributes is ambiguous, so training data inevitably suffer from class imbalance and label sparsity, leading to a lack of effective annotations. An intuitive solution is to find a method to effectively learn image representations by utilizing unlabeled images. With that in mind, we propose a multi-teacher-single-student (MTSS) approach inspired by the multi-task learning and the distillation of semi-supervised learning. Our MTSS learns task-specific domain experts called teacher networks using the label embedding technique and learns a unified model called a student network by forcing a model to mimic the distributions learned by domain experts. Our experiments demonstrate that our method not only achieves competitive performance on various benchmarks for fashion attribute prediction, but also improves robustness and cross-domain adaptability for unseen domains.