Multimodal-aware weakly supervised metric learning with self-weighting triplet loss.

In recent years, we have witnessed a surge of interests in learning distance metrics from various data mining tasks. Most existing methods aim to pull all the similar samples closer while push the dissimilar ones as far as possible. However, when some classes of the dataset exhibit multimodal distribution, these goals conflict and thus can hardly be concurrently satisfied. Additionally, to ensure a valid metric, many methods require a repeated eigenvalue decomposition process, which is time-consuming and numerically unstable. Therefore, how to effectively learn an appropriate distance metric from weakly supervised data remains an open but challenging problem. To address this issue, in this paper, we propose a novel weakly supervised metric learning algorithm, named MultimoDal aware weakly supervised Metric Learning (MDaML). MDaML partitions training data into several clusters and allocates the local cluster center and weight for each sample. Then, combining it with the weighted triplet loss can further enhance the local separability, which encourages local dissimilar samples to keep a large distance from local similar samples. Meanwhile, MDaML casts the metric learning problem into an unconstrained optimization on SPD manifold, which can be efficiently solved by Riemannian Conjugate Gradient Descent (RCGD). Extensive experiments conducted on 13 datasets validate the superiority of the proposed method. [ABSTRACT FROM AUTHOR]