Robust Metric Learning by Smooth Optimization
Huang, Kaizhu, Jin, Rong, Xu, Zenglin, Liu, Cheng-Lin
Most existing distance metric learning methods assume perfect side information that is usually given in pairwise or triplet constraints. Instead, in many real-world applications, the constraints are derived from side information, such as users' implicit feedbacks and citations among articles. As a result, these constraints are usually noisy and contain many mistakes. In this work, we aim to learn a distance metric from noisy constraints by robust optimization in a worst-case scenario, to which we refer as robust metric learning. We formulate the learning task initially as a combinatorial optimization problem, and show that it can be elegantly transformed to a convex programming problem. We present an efficient learning algorithm based on smooth optimization [7]. It has a worst-case convergence rate of O(1/{\surd}{\varepsilon}) for smooth optimization problems, where {\varepsilon} is the desired error of the approximate solution. Finally, our empirical study with UCI data sets demonstrate the effectiveness of the proposed method in comparison to state-of-the-art methods.
Mar-15-2012
- Country:
- Asia (0.29)
- Europe > Germany (0.28)
- North America > United States
- Michigan > Ingham County (0.14)
- Genre:
- Research Report > Promising Solution (0.48)
- Technology: