CVPR 2024 accepted papers on Adversarial robustness and security
63 of the 2,716 papers accepted at CVPR 2024 (2.3%) are on Adversarial robustness and security, compared with 3.2% at CVPR 2023. They fall into 4 topics.
See CVPR 2024 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Attacks on text and graphs · 25 papers
- Physical 3D Adversarial Attacks against Monocular Depth Estimation in Autonomous Driving (54 citations)
- On the Robustness of Large Multimodal Models Against Image Adversarial Attacks (46 citations)
- PAD: Patch-Agnostic Defense against Adversarial Patch Attacks (44 citations)
- One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models (33 citations)
- DAP: A Dynamic Adversarial Patch for Evading Person Detectors (32 citations)
- Watermark-embedded Adversarial Examples for Copyright Protection against Diffusion Models (20 citations)
- Overload: Latency Attacks on Object Detection for Edge Devices (19 citations)
- ASAM: Boosting Segment Anything Model with Adversarial Tuning (18 citations)
- RoDLA: Benchmarking the Robustness of Document Layout Analysis Models (14 citations)
- Language-Driven Anchors for Zero-Shot Adversarial Robustness (10 citations)
- What Sketch Explainability Really Means for Downstream Tasks? (9 citations)
- CosalPure: Learning Concept from Group Images for Robust Co-Saliency Detection (8 citations)
and 13 more in the interactive map.
Certified and empirical robustness · 20 papers
- Pre-trained Model Guided Fine-Tuning for Zero-Shot Adversarial Robustness (27 citations)
- Revisiting Adversarial Training at Scale (27 citations)
- Defense Against Adversarial Attacks on No-Reference Image Quality Models with Gradient Norm Regularization (19 citations)
- Towards Understanding and Improving Adversarial Robustness of Vision Transformers (19 citations)
- MimicDiffusion: Purifying Adversarial Perturbation via Mimicking Clean Diffusion Model (18 citations)
- Revisiting Adversarial Training Under Long-Tailed Distributions (18 citations)
- Robust Distillation via Untargeted and Targeted Intermediate Adversarial Samples (16 citations)
- Robust Overfitting Does Matter: Test-Time Adversarial Purification With FGSM (14 citations)
- Soften to Defend: Towards Adversarial Robustness via Self-Guided Label Refinement (11 citations)
- Adversarial Distillation Based on Slack Matching and Attribution Region Alignment (10 citations)
- Boosting Adversarial Training via Fisher-Rao Norm-based Regularization (10 citations)
- PeerAiD: Improving Adversarial Distillation from a Specialized Peer Tutor (9 citations)
and 8 more in the interactive map.
Backdoor and poisoning · 10 papers
- BadCLIP: Dual-Embedding Guided Backdoor Attack on Multimodal Contrastive Learning (32 citations)
- BadCLIP: Trigger-Aware Prompt Learning for Backdoor Attacks on CLIP (24 citations)
- Adversarial Backdoor Attack by Naturalistic Data Poisoning on Trajectory Prediction in Autonomous Driving (18 citations)
- Data Poisoning based Backdoor Attacks to Contrastive Learning (18 citations)
- Nearest is Not Dearest: Towards Practical Defense against Quantization-conditioned Backdoor Attacks (17 citations)
- LOTUS: Evasive and Resilient Backdoor Attacks through Sub-Partitioning (11 citations)
- Physical Backdoor: Towards Temperature-based Backdoor Attacks in the Physical World (11 citations)
- Not All Prompts Are Secure: A Switchable Backdoor Attack Against Pre-trained Vision Transfomers (9 citations)
- Backdoor Defense via Test-Time Detecting and Repairing (8 citations)
- Attack To Defend: Exploiting Adversarial Attacks for Detecting Poisoned Models (4 citations)
Black-box and transfer attacks · 8 papers
- Boosting Adversarial Transferability by Block Shuffle and Rotation (88 citations)
- DiffAM: Diffusion-based Adversarial Makeup Transfer for Facial Privacy Protection (47 citations)
- Ensemble Diversity Facilitates Adversarial Transferability (34 citations)
- Learning to Transform Dynamically for Better Adversarial Transferability (25 citations)
- Towards Transferable Targeted 3D Adversarial Attack in the Physical World (25 citations)
- Strong Transferable Adversarial Attacks via Ensembled Asymptotically Normal Distribution Learning (19 citations)
- Improving Transferable Targeted Adversarial Attacks with Model Self-Enhancement (15 citations)
- Towards Accurate and Robust Architectures via Neural Architecture Search (11 citations)