AAAI 2025 accepted papers on Adversarial robustness and security
103 of the 3,486 papers accepted at AAAI 2025 (3.0%) are on Adversarial robustness and security, compared with 3.2% at AAAI 2024. They fall into 4 topics.
See AAAI 2025 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 · 44 papers
- A²RNet: Adversarial Attack Resilient Network for Robust Infrared and Visible Image Fusion (19 citations)
- PhishAgent: A Robust Multimodal Agent for Phishing Webpage Detection (18 citations)
- Medical MLLM Is Vulnerable: Cross-Modality Jailbreak and Mismatched Attacks on Medical Multimodal Large Language Models (12 citations)
- MMJ-Bench: A Comprehensive Study on Jailbreak Attacks and Defenses for Vision Language Models (10 citations)
- Medical Multimodal Model Stealing Attacks via Adversarial Domain Alignment (7 citations)
- RAT: Adversarial Attacks on Deep Reinforcement Agents for Targeted Behaviors (7 citations)
- Improving Generalization of Universal Adversarial Perturbation via Dynamic Maximin Optimization (6 citations)
- Sim4Rec: Data-Free Model Extraction Attack on Sequential Recommendation (6 citations)
- Autonomous LLM-Enhanced Adversarial Attack for Text-to-Motion (5 citations)
- Is Your Autonomous Vehicle Safe? Understanding the Threat of Electromagnetic Signal Injection Attacks on Traffic Scene Perception (5 citations)
- KDAT: Inherent Adversarial Robustness via Knowledge Distillation with Adversarial Tuning for Object Detection Models (5 citations)
- NumbOD: A Spatial-Frequency Fusion Attack Against Object Detectors (5 citations)
and 32 more in the interactive map.
Certified and empirical robustness · 28 papers
- AttackBench: Evaluating Gradient-based Attacks for Adversarial Examples (18 citations)
- RP-PGD: Boosting Segmentation Robustness with a Region-and-Prototype Based Adversarial Attack (4 citations)
- SAP: Corrective Machine Unlearning with Scaled Activation Projection for Label Noise Robustness (4 citations)
- 3SAT: A Simple Self-Supervised Adversarial Training Framework (3 citations)
- Can Go AIs Be Adversarially Robust? (3 citations)
- Efficient Robustness Evaluation via Constraint Relaxation (3 citations)
- Enhancing Robustness in Incremental Learning with Adversarial Training (3 citations)
- Enhancing the Adversarial Robustness via Manifold Projection (3 citations)
- Mitigating Feature Gap for Adversarial Robustness by Feature Disentanglement (3 citations)
- AUTE: Peer-Alignment and Self-Unlearning Boost Adversarial Robustness for Training Ensemble Models (2 citations)
- Certified Causal Defense with Generalizable Robustness (2 citations)
- NaFV-Net: An Adversarial Four-view Network for Mammogram Classification (2 citations)
and 16 more in the interactive map.
Backdoor and poisoning · 18 papers
- Adversarial-Inspired Backdoor Defense via Bridging Backdoor and Adversarial Attacks (9 citations)
- UFID: A Unified Framework for Black-box Input-level Backdoor Detection on Diffusion Models (8 citations)
- Watch Out for Your Guidance on Generation! Exploring Conditional Backdoor Attacks against Large Language Models (8 citations)
- CL-Attack: Textual Backdoor Attacks via Cross-Lingual Triggers (5 citations)
- Backdoor Attacks Against No-Reference Image Quality Assessment Models via a Scalable Trigger (4 citations)
- Backdoor Token Unlearning: Exposing and Defending Backdoors in Pretrained Language Models (4 citations)
- Two Sides of the Same Coin: Learning the Backdoor to Remove the Backdoor (4 citations)
- Attack on Prompt: Backdoor Attack in Prompt-Based Continual Learning (3 citations)
- Meme Trojan: Backdoor Attacks Against Hateful Meme Detection via Cross-Modal Triggers (3 citations)
- Speed Master: Quick or Slow Play to Attack Speaker Recognition (3 citations)
- Attention-Imperceptible Backdoor Attacks on Vision Transformers (2 citations)
- Backdoor Attack on Propagation-based Rumor Detectors (2 citations)
and 6 more in the interactive map.
Black-box and transfer attacks · 13 papers
- Improving Integrated Gradient-based Transferable Adversarial Examples by Refining the Integration Path (8 citations)
- Enhancing Adversarial Transferability with Adversarial Weight Tuning (7 citations)
- ADBA: Approximation Decision Boundary Approach for Black-Box Adversarial Attacks (5 citations)
- Query-efficient Attack for Black-box Image Inpainting Forensics via Reinforcement Learning (4 citations)
- Attack-in-the-Chain: Bootstrapping Large Language Models for Attacks Against Black-Box Neural Ranking Models (3 citations)
- Exploring Query Efficient Data Generation Towards Data-Free Model Stealing in Hard Label Setting (2 citations)
- Highly Imperceptible Black-Box Graph Injection Attacks with Reinforcement Learning (2 citations)
- HUANG: A Robust Diffusion Model-based Targeted Adversarial Attack Against Deep Hashing Retrieval (2 citations)
- Reinforcement Learning Platform for Adversarial Black-box Attacks with Custom Distortion Filters (2 citations)
- AIM: Additional Image Guided Generation of Transferable Adversarial Attacks
- Data-Free Universal Attack by Exploiting the Intrinsic Vulnerability of Deep Models
- Everywhere Attack: Attacking Locally and Globally to Boost Targeted Transferability
and 1 more in the interactive map.