Reviews: Generative Local Metric Learning for Kernel Regression

Metric learning is one of the fundamental problems in person re-identification. This paper presents a metric learning method using Nadaraya-Watson (NW) kernel regression. The key feature of the work is that the NW estimator with a learned metric uses information from both the global and local structure of the training data. Theoretical and empirical 9 results confirm that the learned metric can considerably reduce the bias and MSE for kernel regression even when the data are not confined to Gaussian. The main contribution lies in the following aspects: 1. Provided a formation on how metric learning can be embedded in a kernel regression method.