A gradient-type algorithm with backward inertial steps associated to a nonconvex minimization problem.
Cristian Daniel AlecsaSzilárd Csaba LászlóAdrian ViorelPublished in: Numer. Algorithms (2020)
Keyphrases
- objective function
- times faster
- learning algorithm
- computational complexity
- significant improvement
- preprocessing
- detection algorithm
- cost function
- computational cost
- worst case
- recognition algorithm
- kalman filter
- high accuracy
- probabilistic model
- experimental evaluation
- forward and backward
- simulated annealing
- expectation maximization
- convergence rate
- matching algorithm
- forward backward
- decision trees
- global optimization
- convex hull
- optimal solution
- search algorithm
- energy function
- search space
- particle swarm optimization
- dynamic programming
- neural network