adversarial training
A technique in machine learning where a model is trained using both normal and adversarially perturbed inputs. This aims to improve the model's robustness against adversarial attacks and enhances its generalization to unseen data.
- Accelerated Vertical Federated Adversarial Learning through Decoupling Layer-Wise Dependencies
- Autoregressive Adversarial Post-Training for Real-Time Interactive Video Generation
- Bridging Symmetry and Robustness: On the Role of Equivariance in Enhancing Adversarial Robustness
- Distributional Adversarial Attacks and Training in Deep Hedging
- Distributional LLM-as-a-Judge
- Kernel Learning with Adversarial Features: Numerical Efficiency and Adaptive Regularization
- MEIcoder: Decoding Visual Stimuli from Neural Activity by Leveraging Most Exciting Inputs
- MixAT: Combining Continuous and Discrete Adversarial Training for LLMs
- Quantifying Task-relevant Similarities in Representations Using Decision Variable Correlations
- Short-length Adversarial Training Helps LLMs Defend Long-length Jailbreak Attacks: Theoretical and Empirical Evidence
- Solving Neural Min-Max Games: The Role of Architecture, Initialization & Dynamics
- Stealthy Yet Effective: Distribution-Preserving Backdoor Attacks on Graph Classification
- Understanding and Improving Fast Adversarial Training against $l_0$ Bounded Perturbations
- ZEBRA: Towards Zero-Shot Cross-Subject Generalization for Universal Brain Visual Decoding