AdaSTE: An Adaptive Straight-Through Estimator to Train Binary Neural Networks.
Huu LeRasmus Kjær HøierChe-Tsung LinChristopher ZachPublished in: CVPR (2022)
Keyphrases
- neural network
- neural nets
- pattern recognition
- straight line
- fuzzy logic
- back propagation
- least squares
- artificial neural networks
- line segments
- maximum likelihood
- fuzzy systems
- neural network model
- multi class
- fault diagnosis
- maximum a posteriori
- network architecture
- learning algorithm
- self organizing maps
- non binary
- rule extraction
- multi valued
- estimation algorithm
- neural network is trained
- recurrent neural networks
- feed forward
- computational intelligence
- evolutionary algorithm
- pairwise
- genetic algorithm