Revisiting Document Representations for Large-Scale Zero-Shot Learning
This work addresses the challenge of scaling zero-shot learning to large datasets by reducing reliance on human-labeled attributes, though it is incremental in improving document-based representations.
The paper tackles the problem of recognizing unseen objects in large-scale zero-shot learning by proposing a semi-automatic method to extract visual sentences from documents like Wikipedia, reducing human effort. This approach achieves a 64% relative improvement on the ImageNet dataset with over 10,000 unseen classes.
Zero-shot learning aims to recognize unseen objects using their semantic representations. Most existing works use visual attributes labeled by humans, not suitable for large-scale applications. In this paper, we revisit the use of documents as semantic representations. We argue that documents like Wikipedia pages contain rich visual information, which however can easily be buried by the vast amount of non-visual sentences. To address this issue, we propose a semi-automatic mechanism for visual sentence extraction that leverages the document section headers and the clustering structure of visual sentences. The extracted visual sentences, after a novel weighting scheme to distinguish similar classes, essentially form semantic representations like visual attributes but need much less human effort. On the ImageNet dataset with over 10,000 unseen classes, our representations lead to a 64% relative improvement against the commonly used ones.