Training feed-forward networks with the extended Kalman algorithm.
Sharad SinghalLance WuPublished in: ICASSP (1989)
Keyphrases
- filtering algorithm
- computational complexity
- preprocessing
- optimization algorithm
- cost function
- theoretical analysis
- linear programming
- kalman filtering
- training phase
- times faster
- detection algorithm
- segmentation algorithm
- experimental evaluation
- np hard
- optimal solution
- objective function
- reinforcement learning
- expectation maximization
- computationally efficient
- search space
- matching algorithm
- data sets
- computational cost
- genetic algorithm
- recognition algorithm
- convergence rate
- kalman filter
- classification algorithm
- significant improvement
- simulated annealing
- denoising
- supervised learning