An Approach to Stable Gradient Descent Adaptation of Higher-Order Neural Units.
Ivo BukovskyNoriyasu HommaPublished in: CoRR (2016)
Keyphrases
- higher order
- network architecture
- neural network
- high order
- learning rules
- cost function
- natural images
- pairwise
- neural fuzzy
- loss function
- conditional random fields
- associative memory
- artificial neural
- markov random field
- neural model
- adaptation strategies
- connectionist models
- graph cuts
- data sets
- object recognition
- adaptation process
- error function
- bio inspired
- neural computation
- lambda calculus
- nonlinear predictive control