Empirical Evaluation of Active Learning Techniques for Neural MT.
Xiangkai ZengSarthak GargRajen ChatterjeeUdhyakumar NallasamyMatthias PaulikPublished in: DeepLo@EMNLP-IJCNLP (2019)
Keyphrases
- empirical evaluation
- active learning
- network architecture
- machine translation
- neural network
- machine learning
- random sampling
- learning algorithm
- supervised learning
- training examples
- nonlinear predictive control
- pool based active learning
- labeled data
- semi supervised
- learning strategies
- hebbian learning
- experimental design
- bio inspired
- learning process
- neural information processing
- annotation effort
- crowd sourced
- selective sampling
- training set
- associative memory
- relevance feedback
- batch mode active learning
- semi supervised learning
- active learning strategies
- unlabeled data
- biologically inspired
- neural fuzzy
- neural computation
- artificial neural
- support vector
- computational neuroscience
- transfer learning
- sample selection
- reinforcement learning
- spike trains
- neural model
- e learning
- class imbalance
- data sets