Semi-supervised Learning with a Teacher-student Network for Generalized Attribute Prediction

TL;DR

This paper introduces a semi-supervised learning framework using a teacher-student network architecture to improve visual attribute prediction, especially under label sparsity and class imbalance, demonstrating enhanced performance and robustness.

Contribution

The paper proposes a multi-teacher-single-student approach that leverages domain-specific teachers and label embedding for improved semi-supervised attribute prediction.

Findings

Achieves competitive results on fashion attribute benchmarks.

Enhances robustness and cross-domain adaptability.

Utilizes unlabeled images effectively in training.

Abstract

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.