Optimal dimension dependence of the Metropolis-Adjusted Langevin Algorithm.
Sinho ChewiChen LuKwangjun AhnXiang ChengThibaut Le GouicPhilippe RigolletPublished in: CoRR (2020)
Keyphrases
- dynamic programming
- optimal solution
- computational complexity
- learning algorithm
- experimental evaluation
- globally optimal
- preprocessing
- times faster
- worst case
- detection algorithm
- simulated annealing
- computational cost
- np hard
- cost function
- k means
- locally optimal
- theoretical analysis
- improved algorithm
- input data
- search space
- linear programming
- expectation maximization
- state space
- significant improvement
- convex hull
- search algorithm
- objective function
- similarity measure