Exploring Attribute Space with Word Embedding for Zero-shot Learning

With the purpose of addressing the scarcity of attribute diversity in Zero-shot Learning (ZSL), we propose to search for additional attributes in embedding space to extend the class embedding, providing a more discriminative representation of the class prototype. Meanwhile, to tackle the inherent noise behind manually annotated attributes, we apply multi-layer convolutional processing on semantic features rather than conventional linear transformation for filtering. Moreover, we employ Center Loss to assist the training stage, which helps the learned mapping be more accurate and consistent with the corresponding class's prototype. Combining these modules mentioned above, extensive experiments on several public datasets show that our method could yield decent improvements. This proposed way of extending attributes can also be migrated to other models or tasks and obtain better results.

I. Introduction

Existing object recognition methods based on machine learning mainly rely on supervised learning by allowing the model to learn the characteristics of objects through labeled data. According to current research results [1], a larger amount of data generally can bring better results. However, besides the fact that it is unpractical to annotate all the data, the high cost of human labeling and the emerging of new objects must also be considered.