Ridge Regression, Hubness, and Zero-Shot Learning
This addresses hubness issues in zero-shot learning for applications such as cross-lingual and image tasks, though it is incremental as it modifies an existing mapping approach.
The paper tackles hubness in zero-shot learning by proposing to map labels into the example space instead of the reverse, proving and empirically showing that this reduces hubness and improves accuracy in tasks like bilingual lexicon extraction and image labeling.
This paper discusses the effect of hubness in zero-shot learning, when ridge regression is used to find a mapping between the example space to the label space. Contrary to the existing approach, which attempts to find a mapping from the example space to the label space, we show that mapping labels into the example space is desirable to suppress the emergence of hubs in the subsequent nearest neighbor search step. Assuming a simple data model, we prove that the proposed approach indeed reduces hubness. This was verified empirically on the tasks of bilingual lexicon extraction and image labeling: hubness was reduced with both of these tasks and the accuracy was improved accordingly.