后门攻击论文趋势 2023–2026

主要结论

• 后门攻击论文从 2023 年的 73 篇增长到 2026 年的 111 篇，三年间扩大到约 1.5 倍。
• NeurIPS 是发后门攻击论文最多的会议，2023–2026 年共 124 篇，占四个会议总量的 33%。
• 交叉方向里涨得最快的是对齐与安全：同时属于两个主题的论文从 2023 年的 0 篇增至 2026 年的 16 篇。
• 在「安全与可信」的 5 个主题里，它的 2026 年论文数排第 2；同组增长最快的水印是 9.3 倍，后门攻击是 1.5 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026A Theoretical Analysis of Backdoor Learning as Simplicity-Biased Optimization Dynamics
• ICML 2026Coupled Trigger Optimization and Vulnerable Parameter Alignment for Persistent Backdoor Attacks on Federated Learning
• ICLR 2026Defending against Backdoor Attacks via Module Switching
• CVPR 2026Eliminate Distance Differences Induced by Backdoor Attacks Layer-Selective Training and
• NeurIPS 2025A Set of Generalized Components to Achieve Effective Poison-only Clean-label Backdoor Attacks with Collaborative Sample Selection and Triggers
• ICML 2025A Closer Look at Backdoor Attacks on CLIP
• ICLR 2025Activation Gradient based Poisoned Sample Detection Against Backdoor Attacks
• CVPR 2025BadToken: Token-level Backdoor Attacks to Multi-modal Large Language Models
• NeurIPS 2024Reproducibility Study on Adversarial Attacks Against Robust Transformer Trackers
• ICML 2024Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks
• ICLR 2024Adversarial Feature Map Pruning for Backdoor
• CVPR 2024Data Poisoning based Backdoor Attacks to Contrastive Learning
• NeurIPS 2023A Unified Detection Framework for Inference-Stage Backdoor Defenses
• ICML 2023Data Poisoning Attacks Against Multimodal Encoders
• ICLR 2023Adversarial Attacks on Adversarial Bandits
• CVPR 2023Architectural Backdoors in Neural Networks
• NeurIPS 2026AI Models Can Provably Hide Arbitrary Capabilities
• ICML 2026DF-LoGiT: Data-Free Logic-Gated Backdoor Attacks in Vision Transformers
• ICLR 2026Don't Shift the Trigger: Robust Gradient Ascent for Backdoor Unlearning
• CVPR 2026Logit-Margin Repulsion for Backdoor Defense

全部后门攻击论文

共 124 篇，第 1 / 3 页

1. NeurIPS 2026A Theoretical Analysis of Backdoor Learning as Simplicity-Biased Optimization Dynamics
2. NeurIPS 2026AI Models Can Provably Hide Arbitrary Capabilities
3. NeurIPS 2026Backdoor Attacks under Lossy Compression: From Failure to Reactivation and Adaptation
4. NeurIPS 2026Backdoor Channels Hidden in Latent Space: Cryptographic Undetectability in Modern Neural Networks
5. NeurIPS 2026Backdoor Purification for LoRA-Tuned LLMs via Null-Space Projection
6. NeurIPS 2026Beyond Fixed Benchmarks and Worst-Case Attacks: Dynamic Boundary Evaluation for Language Models
7. NeurIPS 2026CASPIAN: Online Detection and Attribution of Cascade Attacks in LLM Multi-Agent Systems via Cross-Channel Causal Monitoring
8. NeurIPS 2026Clean Data Can Still Carry Backdoors: Support-Persistent Backdoors for Model Reuse
9. NeurIPS 2026Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers
10. NeurIPS 2026Consistency-Verified Backdoor Defense for Federated Graph Learning via Cross-Layer Drift
11. NeurIPS 2026DetectViT: Test-time Backdoor Detection for Vision Transformers via Inter-Head Attention Discrepancy
12. NeurIPS 2026Evaluating an Evaluation: Membership Inference Attacks as Machine Unlearning Diagnostics
13. NeurIPS 2026ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?
14. NeurIPS 2026FloatDoor: Platform triggered Backdoors in LLMs
15. NeurIPS 2026HPE: Hallucinated Positive Entanglement for Backdoor Attacks in Federated Self-Supervised Learning
16. NeurIPS 2026Latent-space Attacks for Refusal Evasion in Language Models
17. NeurIPS 2026Learning Robust Representations for Defending White-Box Adversarial Attacks in Continual Learning
18. NeurIPS 2026Leveraging Soft Prompts for Privacy Attacks in Federated Prompt Tuning
19. NeurIPS 2026Local FDR Membership Inference Attacks: Multiple Testing and the Role of Ridge Regularization
20. NeurIPS 2026Neural Reconstruction of LiDAR Point Clouds under Jamming Attacks via Full-Waveform Representation and Simultaneous Laser Sensing
21. NeurIPS 2026Not Suppressing or Purifying: Backdoor Containment via Expert Quarantine and Shutdown in LLMs
22. NeurIPS 2026Poison-then-Hide: Finetuning-Activated Backdoor Attack on Pretrained Vision Encoders
23. NeurIPS 2026Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples
24. NeurIPS 2026Practical Adversarial Attacks on Stochastic Bandits via Fake Data Injection
25. NeurIPS 2026Provable Robustness against Backdoor Attacks via the Primal-Dual Perspective on Differential Privacy
26. NeurIPS 2026Robust and Efficient Backdoor Mitigation for ML Models via Tolerant Property Testing
27. NeurIPS 2026ShadowFPT: Backdooring Federated Prompt Tuning via Shadow Triggers
28. NeurIPS 2026Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks
29. NeurIPS 2026The Platonic Defense: Backdoor Defense for Self-Supervised Encoders in the Era of Large Scale Pre-training
30. NeurIPS 2026Token by Token, Compromised: Backdoor Vulnerabilities in Unified Autoregressive Models
31. NeurIPS 2026Towards Universal Black-box Attacks on Graph Neural Networks
32. NeurIPS 2026TraceGuard: Defending Multi-Turn Jailbreak Attacks via Prompt-Response Risk Signal Tracking
33. NeurIPS 2026Training-Based Backdoors Are Not Cryptographic
34. NeurIPS 2026VOID: Backdoor Injection through Knowledge Vacuity in Federated Unlearning
35. NeurIPS 2026When Sanitization Becomes the Trigger: Defense-Triggered Backdoor Attacks
36. NeurIPS 2026Your Neighbors Know: Leveraging Local Neighborhoods for Backdoor Detection in Decentralized Learning
37. NeurIPS 2025A Set of Generalized Components to Achieve Effective Poison-only Clean-label Backdoor Attacks with Collaborative Sample Selection and Triggers
38. NeurIPS 2025Adversarial Paraphrasing: A Universal Attack for Humanizing AI-Generated Text
39. NeurIPS 2025Attack by Yourself: Effective and Unnoticeable Multi-Category Graph Backdoor Attacks with Subgraph Triggers Pool
40. NeurIPS 2025Backdoor Mitigation via Invertible Pruning Masks
41. NeurIPS 2025BackdoorDM: A Comprehensive Benchmark for Backdoor Learning on Diffusion Model
42. NeurIPS 2025BackdoorLLM: A Comprehensive Benchmark for Backdoor Attacks and Defenses on Large Language Models
43. NeurIPS 2025BadVLA: Towards Backdoor Attacks on Vision-Language-Action Models via Objective-Decoupled Optimization
44. NeurIPS 2025Competitive Advantage Attacks to Decentralized Federated Learning
45. NeurIPS 2025Defending Multimodal Backdoored Models by Repulsive Visual Prompt Tuning
46. NeurIPS 2025Efficient Federated Learning against Byzantine Attacks and Data Heterogeneity via Aggregating Normalized Gradients
47. NeurIPS 2025Enhancing LLM Watermark Resilience Against Both Scrubbing and Spoofing Attacks
48. NeurIPS 2025FedRACE: A Hierarchical and Statistical Framework for Robust Federated Learning
49. NeurIPS 2025From Dormant to Deleted: Tamper-Resistant Unlearning Through Weight-Space Regularization
50. NeurIPS 2025ICLScan: Detecting Backdoors in Black-Box Large Language Models via Targeted In-context Illumination

常见问题

2023–2026 年四大会议一共有多少篇后门攻击论文？

共 381 篇，按年份分别是 2023 年 73 篇、2024 年 105 篇、2025 年 92 篇、2026 年 111 篇。

NeurIPS 2026 有多少篇后门攻击论文？

NeurIPS 2026 有 36 篇后门攻击论文，上一届是 31 篇。

ICLR 2026 有多少篇后门攻击论文？

ICLR 2026 有 20 篇后门攻击论文，上一届是 25 篇。

ICML 2026 有多少篇后门攻击论文？

ICML 2026 有 36 篇后门攻击论文，上一届是 17 篇。

CVPR 2026 有多少篇后门攻击论文？

CVPR 2026 有 19 篇后门攻击论文，上一届是 19 篇。

哪个会议的后门攻击论文最多？

NeurIPS，2023–2026 年共 124 篇。