Abstract

Relative attributes learning aims to learn ranking func-tions describing the relative strength of attributes. Mostof current learning approaches learn ranking functions foreach attribute independently without considering possibleintrinsic relatedness among the attributes. For a problem involving multiple attributes, it is reasonable to assume that utilizing such relatedness among the attributes would benefit learning, especially when the number of labeled training pairs are very limited. In this paper, we proposed a relative multi-attribute learning framework that integrates relative attributes into a multi-task learning scheme. Theformulation allows us to exploit the advantages of the state-of-the-art regularization-based multi-task learning for im-proved attribute learning. In particular, using joint feature learning as the case studies, we evaluated our framework with both synthetic data and two real datasets. Experimental results suggest that the proposed framework has clear performance gain in ranking accuracy and zero-shot learning accuracy over existing methods of independent relative attributes learning and multi-task learning.