FLONet: Fewer Labeling Cost Active Learning for Deep Neural Network.
Xianya FuRui WangPeixuan ZuoJiaming ZhouJia ZhaiXiaodan XieZhongzhi LuanDepei QianPublished in: HPCC/SmartCity/DSS (2019)
Keyphrases
- active learning
- neural network
- active learning strategies
- cost sensitive
- artificial neural networks
- cost sensitive learning
- sample selection
- random sampling
- selective sampling
- learning process
- semi supervised
- active sampling
- machine learning
- training examples
- training set
- learning strategies
- fuzzy logic
- genetic algorithm
- labeling effort
- neural network model
- fault diagnosis
- supervised learning
- random selection
- batch mode
- selection strategies
- prediction model
- high cost
- experimental design
- minimum cost
- hidden layer
- total cost
- multi layer
- bp neural network
- data sets
- back propagation
- unlabeled data
- labeled data
- relevance feedback
- network architecture
- learning vector quantization
- expected loss
- neural network is trained
- pairwise
- pattern recognition
- query by committee