New and improved version of LMNN. Now runs more stable and has much fewer dependencies. Should be faster on multi-core machines. Now with pre-compiled binaries for Linux, Mac and Windows. (This code was uploaded on 09/18/2012, please let me know if you run into any problems installing or running it. Thanks.)

Code for Large Margin Nearest Neighbors

Maximum Variance Unfolding Matlab Code (original code + landmark version) [Previously called Semidefinite Embedding (SDE)] This code contains the landmark MVU version (AISTATS'05), the Graph Laplacien Regularized version (NIPS'06) and the original MVU code (IJCV'05) (See also http://homepage.tudelft.nl/19j49/Matlab_Toolbox_for_Dimensionality_Reduction.html )

Fast C++ implementation of Gradient Boosted Regression Trees and Random Forests (by Ananth Mohan)

Multi-Task LMNN (NIPS 2010) [Code by Shibin Parameswaran]

High performance parallel implementation (C++/MPI) of gradient boosted regression trees.

marginalized Stacked Denoising Autoencoder (mSDA)

Re-implementation of Metric learning for kernel regression (MLKR).

Learnign with Marginalized Corrupted Features (Matlab implementation by Laurens van der Maaten).