Label Embedding Network: Learning Label Representation for Soft Training of Deep Networks

TL;DR

This paper introduces a Label Embedding Network that learns label representations during training, transforming the loss function into a soft distribution to improve accuracy, convergence speed, and interpretability in deep learning models.

Contribution

It presents a novel method for automatically learning label embeddings during training, enhancing model performance and interpretability over traditional one-hot label representations.

Findings

Achieves higher accuracy than state-of-the-art methods

Faster convergence in training deep networks

Learned label embeddings are reasonable and interpretable

Abstract

We propose a method, called Label Embedding Network, which can learn label representation (label embedding) during the training process of deep networks. With the proposed method, the label embedding is adaptively and automatically learned through back propagation. The original one-hot represented loss function is converted into a new loss function with soft distributions, such that the originally unrelated labels have continuous interactions with each other during the training process. As a result, the trained model can achieve substantially higher accuracy and with faster convergence speed. Experimental results based on competitive tasks demonstrate the effectiveness of the proposed method, and the learned label embedding is reasonable and interpretable. The proposed method achieves comparable or even better results than the state-of-the-art systems. The source code is available at \url{https://github.com/lancopku/LabelEmb}.