The core code of the paper "Towards Certified Robustness of Distance Metric Learning".

• The file demo.m shows how to implement the LMNN-PL method.
• The file LMNN_PL.m file is the main function.

• Our code borrows some helper functions from the packages of Large Margin Nearest Neighbor (JMLR09) and Sparse Compositional Metric Learning (AAAI14). We thanks the writers a lot.

• v1 (18/09/2020): initial release.

If you have any queries about the code, please feel free to contact me: Xiaochen Yang (xiaochen.yang.16@ucl.ac.uk)