Boosting the Performance of Generic Deep Neural Network Frameworks with Log-supermodular CRFs.
Hao XiongYangxiao LuNicholas RuozziPublished in: NeurIPS (2022)
Keyphrases
- neural network
- conditional random fields
- higher order
- back propagation
- artificial neural networks
- weighted sums
- neural network model
- learning algorithm
- unsupervised feature learning
- neural network is trained
- fully connected
- machine learning
- partially labeled data
- feature selection
- pattern recognition
- fuzzy logic
- radial basis function
- activation function
- graphical models
- named entity recognition
- network architecture
- hidden markov models
- sequence labeling
- boosting algorithms
- maximum entropy
- ensemble learning
- bp neural network
- multilayer perceptron
- weak learners
- knn
- crf model
- feed forward
- classification and regression trees
- probabilistic model
- information extraction
- self organizing maps
- multi class learning
- fault diagnosis