近期论文23
- Can Small Language Models Reliably Resist Jailbreak Attacks? A Comprehensive Evaluation第 1 位,共 6 位作者CCS 2026 · 11 次引用
- PromptCOS: Towards Content-Only System Prompt Copyright Auditing for LLMs第 7 位,共 9 位作者S&P 2026 · 5 次引用
- JailbreakScope: Interpreting Jailbreak Mechanism through Representation and Circuit Analyses第 2 位,共 7 位作者USENIX Security 2026
- On Evaluating the Robustness of Large Vision-Language Models via Untargeted Modality Alignment Breaking Adversarial Attack第 3 位,共 8 位作者USENIX Security 2026
- ICLScan: Detecting Backdoors in Black-Box Large Language Models via Targeted In-context Illumination第 5 位,共 5 位作者NeurIPS 2025 · 7 次引用
- PoiSAFL: Scalable Poisoning Attack Framework to Byzantine-resilient Semi-asynchronous Federated Learning第 3 位,共 9 位作者USENIX Security 2025
- SoK: On Gradient Leakage in Federated Learning第 3 位,共 8 位作者USENIX Security 2025
- Textual Unlearning Gives a False Sense of Unlearning第 2 位,共 6 位作者ICML 2025
- Byzantine-robust Decentralized Federated Learning via Dual-domain Clustering and Trust Bootstrapping第 3 位,共 4 位作者CVPR 2024 · 21 次引用
- Breaking Secure Aggregation: Label Leakage from Aggregated Gradients in Federated Learning第 1 位,共 7 位作者INFOCOM 2024 · 10 次引用
- FaceObfuscator: Defending Deep Learning-based Privacy Attacks with Gradient Descent-resistant Features in Face Recognition第 3 位,共 11 位作者USENIX Security 2024 · 9 次引用
- A Causal Explainable Guardrails for Large Language Models第 4 位,共 6 位作者CCS 2024 · 5 次引用
- DAAP: Privacy-Preserving Model Accuracy Estimation on Unlabeled Datasets Through Distribution-Aware Adversarial Perturbation第 2 位,共 4 位作者USENIX Security 2024
- Locate and Verify: A Two-Stream Network for Improved Deepfake Detection第 5 位,共 9 位作者ACM MM 2023 · 52 次引用
- DFIL: Deepfake Incremental Learning by Exploiting Domain-invariant Forgery Clues第 7 位,共 9 位作者ACM MM 2023 · 35 次引用
- Counterfactual-based Saliency Map: Towards Visual Contrastive Explanations for Neural Networks第 2 位,共 8 位作者ICCV 2023 · 15 次引用
- Towards Fairness-aware Adversarial Network Pruning第 2 位,共 5 位作者ICCV 2023 · 8 次引用
- Privacy-preserving Adversarial Facial Features第 1 位,共 10 位作者CVPR 2023
- Towards Transferable Targeted Adversarial Examples第 1 位,共 7 位作者CVPR 2023
- mmEve: eavesdropping on smartphone's earpiece via COTS mmWave device第 5 位,共 9 位作者MobiCom 2022 · 60 次引用
- Fairness-aware Adversarial Perturbation Towards Bias Mitigation for Deployed Deep Models第 1 位,共 6 位作者CVPR 2022 · 49 次引用
- "Adversarial Examples" for Proof-of-Learning第 4 位,共 6 位作者S&P 2022 · 41 次引用
- Feature Importance-aware Transferable Adversarial Attacks第 1 位,共 6 位作者ICCV 2021 · 306 次引用
