Flooding-X: Improving BERT’s Resistance to Adversarial Attacks via Loss-Restricted Fine-Tuning

Flooding-X: Improving BERT’s Resistance to Adversarial Attacks via Loss-Restricted Fine-Tuning

RobustLearning2022ACLadversarialheuristicnovelsmoothingflooding
Do We Need Zero Training Loss After Achieving Zero Training Error?

Do We Need Zero Training Loss After Achieving Zero Training Error?

NeuralNetworks2020emmmheuristicICMLoverfittingsmoothing

标签