SPINDLE: Self-Pretrainable In-situ Normalizer for Deep Learning Error Function.
Jyun-Xin KeSiang-En CiouTing-Yu ChenTsung-Chu HuangPublished in: ICCE-TW (2022)
Keyphrases
- deep learning
- error function
- learning rate
- unsupervised learning
- feedforward neural networks
- activation function
- machine learning
- cost function
- fuzzy neural network
- global minimum
- convergence rate
- linear regression
- weakly supervised
- mental models
- neural network
- basis functions
- high dimensional
- decision making
- learning algorithm