Towards Reliable Convergence in the Training of Neural Networks - The Streamlined Glide Algorithm and the LM Glide Algorithm.
Vitit KantabutraPublished in: MLMTA (2006)
Keyphrases
- neural network
- learning algorithm
- convergence rate
- objective function
- high accuracy
- computational cost
- cost function
- experimental evaluation
- preprocessing
- improved algorithm
- np hard
- dynamic programming
- image segmentation
- search space
- optimization algorithm
- times faster
- computationally efficient
- convergence property
- training process
- classification algorithm
- clustering method
- detection algorithm
- expectation maximization
- data sets
- k means
- convergence speed
- particle swarm optimization