PILE: Pairwise Iterative Logits Ensemble for Multi-Teacher Labeled Distillation.
Lianshang CaiLinhao ZhangDehong MaJun FanDaiting ShiYi WuZhicong ChengSimiu GuDawei YinPublished in: EMNLP (Industry Track) (2022)
Keyphrases
- pairwise
- training data
- supervised learning
- semi supervised
- neural network
- loss function
- training set
- similarity measure
- markov random field
- ensemble learning
- data sets
- higher order
- random forests
- multi class
- decision trees
- feature selection
- semi supervised learning
- random forest
- professional development
- iterative methods
- pairwise interactions