Implement one of the following matrix decompositions 1. UV Decomposition using stochastic gradient descent 2. randomized SVD 3. matrix sketching using frequent directions 4. CUR Decomposition You are free to choose any of the above algorithms. You also have full freedom to implement these in any framework. Potential choices are, 1. MPI 2. Hadoop 3. Spark Any other distributed computing framework is also acceptable. Please check with me before starting implementation using any other framework. You should use any standard library for computing the sequential SVD required for some of these algorithms. As always, commit your code often to github and submit a PDF report describing your experience and evaluation of your implementation.