AAAI 2024 accepted papers on Adversarial robustness and security
93 of the 2,864 papers accepted at AAAI 2024 (3.2%) are on Adversarial robustness and security, compared with 3.5% at AAAI 2023. They fall into 4 topics.
See AAAI 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 · 34 papers
- Visual Adversarial Examples Jailbreak Aligned Large Language Models (122 citations)
- Learning to Unlearn: Instance-Wise Unlearning for Pre-trained Classifiers (40 citations)
- Adv-Diffusion: Imperceptible Adversarial Face Identity Attack via Latent Diffusion Model (35 citations)
- Beyond Mimicking Under-Represented Emotions: Deep Data Augmentation with Emotional Subspace Constraints for EEG-Based Emotion Recognition (27 citations)
- SlowTrack: Increasing the Latency of Camera-Based Perception in Autonomous Driving Using Adversarial Examples (22 citations)
- Enhancing the Robustness of Spiking Neural Networks with Stochastic Gating Mechanisms (13 citations)
- Towards Robust Image Stitching: An Adaptive Resistance Learning against Compatible Attacks (13 citations)
- Comparing the Robustness of Modern No-Reference Image- and Video-Quality Metrics to Adversarial Attacks (11 citations)
- Stable Unlearnable Example: Enhancing the Robustness of Unlearnable Examples via Stable Error-Minimizing Noise (11 citations)
- Data-Free Hard-Label Robustness Stealing Attack (10 citations)
- MathAttack: Attacking Large Language Models towards Math Solving Ability (10 citations)
- LimeAttack: Local Explainable Method for Textual Hard-Label Adversarial Attack (9 citations)
and 22 more in the interactive map.
Certified and empirical robustness · 26 papers
- Sparse Enhanced Network: An Adversarial Generation Method for Robust Augmentation in Sequential Recommendation (18 citations)
- Taxonomy Driven Fast Adversarial Training (17 citations)
- Improving the Adversarial Transferability of Vision Transformers with Virtual Dense Connection (16 citations)
- Adversarial Initialization with Universal Adversarial Perturbation: A New Approach to Fast Adversarial Training (12 citations)
- Out of Thin Air: Exploring Data-Free Adversarial Robustness Distillation (9 citations)
- Perturbation-Invariant Adversarial Training for Neural Ranking Models: Improving the Effectiveness-Robustness Trade-Off (8 citations)
- Promoting Counterfactual Robustness through Diversity (8 citations)
- Adversarial Purification with the Manifold Hypothesis (7 citations)
- Harnessing Edge Information for Improved Robustness in Vision Transformers (7 citations)
- Representation-Based Robustness in Goal-Conditioned Reinforcement Learning (6 citations)
- Robust Few-Shot Named Entity Recognition with Boundary Discrimination and Correlation Purification (6 citations)
- DRF: Improving Certified Robustness via Distributional Robustness Framework (5 citations)
and 14 more in the interactive map.
Backdoor and poisoning · 17 papers
- Beyond Traditional Threats: A Persistent Backdoor Attack on Federated Learning (41 citations)
- Backdoor Attacks via Machine Unlearning (28 citations)
- Elijah: Eliminating Backdoors Injected in Diffusion Models via Distribution Shift (27 citations)
- Personalization as a Shortcut for Few-Shot Backdoor Attack against Text-to-Image Diffusion Models (23 citations)
- COMBAT: Alternated Training for Effective Clean-Label Backdoor Attacks (21 citations)
- Conditional Backdoor Attack via JPEG Compression (14 citations)
- DataElixir: Purifying Poisoned Dataset to Mitigate Backdoor Attacks via Diffusion Models (14 citations)
- Progressive Poisoned Data Isolation for Training-Time Backdoor Defense (14 citations)
- Chronic Poisoning: Backdoor Attack against Split Learning (12 citations)
- BadRL: Sparse Targeted Backdoor Attack against Reinforcement Learning (11 citations)
- Does Few-Shot Learning Suffer from Backdoor Attacks? (11 citations)
- Inspecting Prediction Confidence for Detecting Black-Box Backdoor Attacks (10 citations)
and 5 more in the interactive map.
Black-box and transfer attacks · 16 papers
- Boosting Adversarial Transferability across Model Genus by Deformation-Constrained Warping (25 citations)
- DifAttack: Query-Efficient Black-Box Adversarial Attack via Disentangled Feature Space (22 citations)
- Once and for All: Universal Transferable Adversarial Perturbation against Deep Hashing-Based Facial Image Retrieval (17 citations)
- AGS: Affordable and Generalizable Substitute Training for Transferable Adversarial Attack (14 citations)
- LRS: Enhancing Adversarial Transferability through Lipschitz Regularized Surrogate (12 citations)
- On the Convergence of an Adaptive Momentum Method for Adversarial Attacks (12 citations)
- Towards Transferable Adversarial Attacks with Centralized Perturbation (11 citations)
- Transferable Adversarial Attacks for Object Detection Using Object-Aware Significant Feature Distortion (11 citations)
- FACL-Attack: Frequency-Aware Contrastive Learning for Transferable Adversarial Attacks (9 citations)
- Mutual-Modality Adversarial Attack with Semantic Perturbation (9 citations)
- Attacking Transformers with Feature Diversity Adversarial Perturbation (8 citations)
- TTTS: Tree Test Time Simulation for Enhancing Decision Tree Robustness against Adversarial Examples (8 citations)
and 4 more in the interactive map.