Balancing Robustness and Sensitivity using Feature Contrastive Learning
It is generally believed that robust training of extremely large networks is critical to their success in real-world applications. However, when taken to the extreme, methods that promote robustness can hurt the model's sensitivity to rare or underrepresented patterns. In this paper, we discuss this trade-off between sensitivity and robustness to natural (non-adversarial) perturbations by introducing two notions: contextual feature utility and contextual feature sensitivity. We propose Feature Contrastive Learning (FCL) that encourages a model to be more sensitive to the features that have higher contextual utility. Empirical results demonstrate that models trained with FCL achieve a better balance of robustness and sensitivity, leading to improved generalization in the presence of noise on both vision and NLP datasets.
Code (0)
등록된 구현이 없습니다.
Tasks
Contrastive LearningSensitivityMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Balanced Adversarial Training: Balancing Tradeoffs Between Oversensitivity and Undersensitivity in NLP Models
Traditional (\emph{oversensitive}) adversarial examples involve finding a small perturbation that does not change an input's true label but confuses the classifier into outputting a different prediction. \emph{Undersensi…
Contrastive LearningNatural Language InferenceParaphrase IdentificationUnderstanding and Improving Training-Free AI-Generated Image Detections with Vision Foundation Models
The rapid advancement of generative models has introduced serious risks, including deepfake techniques for facial synthesis and editing. Traditional approaches rely on training classifiers and enhancing generalizability …
DeepFake DetectionFace SwappingSensitivityPareto Adversarial Robustness: Balancing Spatial Robustness and Sensitivity-based Robustness
Adversarial robustness, which primarily comprises sensitivity-based robustness and spatial robustness, plays an integral part in achieving robust generalization. In this paper, we endeavor to design strategies to achieve…
Adversarial RobustnessSensitivityAdversarial Feature Alignment: Balancing Robustness and Accuracy in Deep Learning via Adversarial Training
Deep learning models continue to advance in accuracy, yet they remain vulnerable to adversarial attacks, which often lead to the misclassification of adversarial examples. Adversarial training is used to mitigate this pr…
Contrastive LearningData AugmentationLearning Time-Series Representations by Hierarchical Uniformity-Tolerance Latent Balancing
We propose TimeHUT, a novel method for learning time-series representations by hierarchical uniformity-tolerance balancing of contrastive representations. Our method uses two distinct losses to learn strong representatio…
Temporal SequencesAnomaly Detection