Robust Knowledge Distillation Based on Feature Variance Against Backdoored Teacher Model.
Jinyin ChenXiaoming ZhaoHaibin ZhengXiao LiSheng XiangHaifeng GuoPublished in: CoRR (2024)
Keyphrases
- experimental data
- prior knowledge
- high level
- conceptual model
- computational model
- machine learning
- formal model
- statistical model
- probabilistic model
- domain experts
- semantic models
- expert knowledge
- mathematical model
- domain knowledge
- feature vectors
- image segmentation
- theoretical framework
- knowledge management
- objective function
- correlation coefficient
- reinforcement learning
- bayesian networks