We present a Lagrangian decomposition approach for the Knapsack Constrained Maximum Spanning Tree problem yielding upper bounds as well as heuristic solutions. This method is furth...
Genetic algorithms are a population-based Meta heuristics. They have been successfully applied to many optimization problems. However, premature convergence is an inherent charact...
— A novel sparse kernel density estimator is derived based on a regression approach, which selects a very small subset of significant kernels by means of the D-optimality experi...
In least squares support vector (LS-SVM), the key challenge lies in the selection of free parameters such as kernel parameters and tradeoff parameter. However, when a large number ...
Many applications require optimizing an unknown, noisy function that is expensive to evaluate. We formalize this task as a multiarmed bandit problem, where the payoff function is ...
Niranjan Srinivas, Andreas Krause, Sham Kakade, Ma...