A Poisoning Attack Against the Recognition Model Trained by the Data Augmentation Method.
Yunhao YangLong LiLiang ChangTianlong GuPublished in: ML4CS (2) (2020)
Keyphrases
- input data
- test data
- synthetic data
- prior knowledge
- prior information
- statistical methods
- probabilistic model
- recognition algorithm
- data sets
- objective function
- graph representation
- noisy data
- experimental data
- statistical model
- dimensional vector
- recognition process
- cost function
- theoretical analysis
- training procedure
- training data
- predictive model
- linear model
- probability distribution
- prediction model
- em algorithm
- mathematical model
- similarity measure
- estimation process
- training samples
- pairwise
- missing data
- hidden markov models
- recognition rate
- gaussian distribution
- statistical information
- regression analysis
- learned models
- energy function
- missing values
- data sources
- data points
- network structure