Stochastic (block) BFGS method for solving the empirical risk minimization problem with  logistic loss and L2 regularizer. Related paper.
Random Inverse
A suite of randomized methods for inverting positive definite matrices implemented in MATLAB. Related paper.
Random Linear Lab
A lab for testing and comparing randomized methods for solving linear systems. Implemented in MATLAB. Related paper.
A framework for communication-efficient distributed optimization for machine learning.

Accelerated, Parallel and PROXimal coordinate descent. This is an efficient C++ code based on this paper. We also implement PCDM (parallel coordinate descent), SDCA (stochastic dual coordinate ascent) and AGD (Accelerated Gradient Descent).
APPROX is implemented in APPROX @ scikit-learn. Choose option "accelerated" when doing LASSO/elastic net.


Semi-stochastic gradient descent method for fast training of L2 regularized logistic regression. This is an efficient C++ code (can be called from MATLAB), based on this paper.

Parallel Sparse PCA [8 9] code. Supports multicore workstations, GPUs and clusters. The cluster version was tested on terabyte matrices and is scalable. Extension of GPower.

Serial [1 5], parallel [2 3 4] and distributed [6 7] coordinate descent code for big data optimization. The parallel and distributed codes can solve LASSO instances with terabyte matrices and billions of features, and are scalable.