NIPS 2017 - presentations from the Optimization session

• On the Optimization Landscape of Tensor Decompositions • Robust Optimization for Non-Convex Objectives • Bayesian Optimization with Gradients • Gradient Descent Can Take Exponential Time to Escape Saddle Points • Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration • Limitations on Variance-Reduction and Acceleration Schemes for Finite Sums Optimization • Implicit Regularization in Matrix Factorization • Linear Convergence of a Frank-Wolfe Type Algorithm over Trace-Norm Balls • Acceleration and Averaging in Stochastic Descent Dynamics When Cyclic Coordinate Descent Beats Randomized Coordinate Descent