Backdoor Attacks papers, 2023–2026

Key findings

• Backdoor Attacks went from 73 papers in 2023 to 111 in 2026, roughly 1.5x in three years.
• NeurIPS publishes the most Backdoor Attacks work: 124 papers from 2023 to 2026, 33% of the four-venue total.
• The fastest-growing overlap is with Alignment and Safety: papers tagged with both rose from 0 in 2023 to 16 in 2026.
• Among the 5 topics under Security and Trust, it ranks #2 by 2026 paper count; the fastest grower in the group, Watermarking, is up 9.3x versus 1.5x here.

Representative papers

Spread across years and venues so every edition shows up

• 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

All Backdoor Attacks papers

36 papers, page 1 of 1

1. ICML 2026Angel or Demon: Investigating the Plasticity Interventions' Impact on Backdoor Threats in Deep Reinforcement Learning
2. ICML 2026Anti-Backdoor Coreset Selection via Cumulative Entropy
3. ICML 2026Attention Hijacking: Backdooring Text Dataset Distillation via Semantic Anchors
4. ICML 2026BOCLOAK: Optimal Transport-Guided Adversarial Attacks on Graph Neural Network-Based Bot Detection
5. ICML 2026BYORn: Bootstrap Your Own Responses to Defend Large Vision-Language Models Against Backdoor Attacks
6. ICML 2026CBV: Clean-label Backdoor Attacks on Vision Language Models via Diffusion Models
7. ICML 2026Coupled Trigger Optimization and Vulnerable Parameter Alignment for Persistent Backdoor Attacks on Federated Learning
8. ICML 2026DF-LoGiT: Data-Free Logic-Gated Backdoor Attacks in Vision Transformers
9. ICML 2026Erased but Not Forgotten: How Backdoors Compromise Concept Erasure
10. ICML 2026Fox in the Henhouse: Supply-Chain Backdoor Attacks Against Reinforcement Learning
11. ICML 2026From Internal Diagnosis to External Auditing: A VLM-Driven Paradigm for Data-Free Online Backdoor Defense
12. ICML 2026From Parameters to Feature Space: Task Arithmetic for Backdoor Mitigation in Model Merging
13. ICML 2026From Poisoned to Aware: Fostering Backdoor Self-Awareness in LLMs
14. ICML 2026Improving the Sensitivity of Backdoor Detectors via Class Subspace Orthogonalization
15. ICML 2026INDEXGUARD: Index-only Backdoor Vetting for Secure Federated PEFT of Large Language Models
16. ICML 2026Neural Honeytrace: Plug&Play Watermarking Framework against Model Extraction Attacks
17. ICML 2026New Wide-Net-Casting Jailbreak Attacks Risk Large Models
18. ICML 2026PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks
19. ICML 2026Persistent Backdoor Attacks in Class-Incremental Learning via Structural Invariant Anchoring
20. ICML 2026PolicyGuard: Towards Test-time and Step-level Adversary Defense for Reinforcement Learning Agent
21. ICML 2026Rapid Poison: Practical Poisoning Attacks Against the Rapid Response Framework
22. ICML 2026Rethinking Forgery Attacks on Semantic Watermarks in Black-Box Settings: A Geometric Distortion Perspective
23. ICML 2026Robust Harmful Features Under Jailbreak Attacks: Mechanistic Evidence from Attention Head Specialization in Large Language Models
24. ICML 2026Robust In-Context Reinforcement Learning Under Reward Poisoning Attacks
25. ICML 2026Semantic-level Backdoor Attack against Text-to-Image Diffusion Models
26. ICML 2026SemBind: Binding Diffusion Watermarks to Semantics Against Black-Box Forgery Attacks
27. ICML 2026SkillTrojan: Backdoor Attacks on Skill-Based Agent Systems
28. ICML 2026SORA: Free Second-Order Attacks in Fast Adversarial Training
29. ICML 2026TCAP: Tri-Component Attention Profiling for Unsupervised Backdoor Detection in MLLM Fine-Tuning
30. ICML 2026Theory of Continual Learning Against Data Poisoning Attacks
31. ICML 2026Theory of Minimal Weight Perturbations in Deep Networks and its Applications for Low-Rank Activated Backdoor Attacks
32. ICML 2026TimeGuard: Channel-wise Pool Training for Backdoor Defense in Time Series Forecasting
33. ICML 2026TokenSwap: Backdoor Attack on the Compositional Understanding of Large Vision-Language Models
34. ICML 2026Toward Subspace-Perturbed Trajectory-Aware Backdoor Attacks in Deep Reinforcement Learning
35. ICML 2026Turning Bias into Bugs: Bandit-Guided Style Manipulation Attacks on LLM Judges
36. ICML 2026Uncovering Hidden Triggers: Backdoor Attribution in Language Models

FAQ

How many Backdoor Attacks papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

381 in total: 73 in 2023, 105 in 2024, 92 in 2025, 111 in 2026.

How many Backdoor Attacks papers are at NeurIPS 2026?

NeurIPS 2026 has 36 Backdoor Attacks papers, versus 31 the year before.

How many Backdoor Attacks papers are at ICLR 2026?

ICLR 2026 has 20 Backdoor Attacks papers, versus 25 the year before.

How many Backdoor Attacks papers are at ICML 2026?

ICML 2026 has 36 Backdoor Attacks papers, versus 17 the year before.

How many Backdoor Attacks papers are at CVPR 2026?

CVPR 2026 has 19 Backdoor Attacks papers, versus 19 the year before.

Which conference publishes the most Backdoor Attacks papers?

NeurIPS, with 124 papers from 2023 to 2026.