An Adversarial Training Method for Improving Model Robustness in Unsupervised Domain Adaptation.
Zhishen NieYing LinMeng YanYifan CaoShengfu NingPublished in: KSEM (2021)
Keyphrases
- mathematical model
- probabilistic model
- objective function
- statistical model
- cost function
- linear model
- test data
- study proposes
- evaluation method
- prediction model
- hybrid model
- significant improvement
- training process
- theoretical analysis
- training samples
- network model
- clustering method
- modeling method
- tree structure
- pairwise
- similarity measure
- classification method
- preprocessing
- monte carlo simulation
- training set
- sensitivity analysis
- statistical methods
- em algorithm
- prior knowledge
- segmentation method
- training algorithm
- regression analysis
- training phase
- closed form
- bp neural network
- bayesian framework
- prior information
- classification algorithm
- energy function
- parameter estimation
- input data
- recognition algorithm
- parameter space
- optimization method
- hybrid method
- evaluation model
- reconstruction method
- dynamic programming
- training procedure