Label Propagation for Deep Semi-supervised Learning
This work addresses the problem of leveraging unlabeled data for deep learning, offering a complementary approach to current state-of-the-art methods, though it is incremental as it adapts classic transductive techniques to modern inductive frameworks.
The paper tackles the challenge of semi-supervised learning by integrating a transductive label propagation method based on the manifold assumption to generate pseudo-labels for unlabeled data, training a deep neural network iteratively, resulting in improved performance on several datasets, particularly in the few-labels regime.
Semi-supervised learning is becoming increasingly important because it can combine data carefully labeled by humans with abundant unlabeled data to train deep neural networks. Classic methods on semi-supervised learning that have focused on transductive learning have not been fully exploited in the inductive framework followed by modern deep learning. The same holds for the manifold assumption---that similar examples should get the same prediction. In this work, we employ a transductive label propagation method that is based on the manifold assumption to make predictions on the entire dataset and use these predictions to generate pseudo-labels for the unlabeled data and train a deep neural network. At the core of the transductive method lies a nearest neighbor graph of the dataset that we create based on the embeddings of the same network.Therefore our learning process iterates between these two steps. We improve performance on several datasets especially in the few labels regime and show that our work is complementary to current state of the art.