对齐与安全论文趋势 2023–2026

主要结论

• 对齐与安全论文从 2023 年的 438 篇增长到 2026 年的 1,882 篇，三年间扩大到约 4.3 倍。
• NeurIPS 是发对齐与安全论文最多的会议，2023–2026 年共 1,421 篇，占四个会议总量的 31%。
• 交叉方向里涨得最快的是社会层面 · safety：同时属于两个主题的论文从 2023 年的 0 篇增至 2026 年的 137 篇。
• 在「对齐与负责任 AI」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的社会层面 · safety是 93 倍，对齐与安全是 4.3 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 2026$\gamma$-weakly $\theta$-up-concavity: A Unified Framework for Non-Convex Optimization Beyond DR-Submodular and OSS Functions
• ICML 2026AICrypto: Evaluating Cryptography Capabilities of Large Language Models
• ICLR 2026A Fair Bayesian Inference through Matched Gibbs Posterior
• CVPR 2026Advancing Cancer Prognosis with Hierarchical Fusion of Genomic Proteomic and
• NeurIPS 2025$\Psi$-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Models
• ICML 2025A Checks-and-Balances Framework for Context-Aware Ethical AI Alignment
• ICLR 2025$\sigma$-zero: Gradient-based Optimization of $\ell_0$-norm Adversarial Examples
• CVPR 2025AesthetiQ: Enhancing Graphic Layout Design via Aesthetic-Aware Preference Alignment of Multi-modal Large Language Models
• NeurIPS 2024A survey and benchmark of high-dimensional Bayesian optimization of discrete sequences
• ICML 2024A Global Geometric Analysis of Maximal Coding Rate Reduction
• ICLR 2024A unique M-pattern for micro-expression spotting in long videos
• CVPR 2024Decompose-and-Compose: A Compositional Approach to Mitigating Spurious Correlation
• NeurIPS 2023(S)GD over Diagonal Linear Networks: Implicit bias, Large Stepsizes and Edge of Stability
• ICML 2023A Robust Optimisation Perspective on Counterexample-Guided Repair of Neural Networks
• ICLR 2023Aligning Model and Macaque Inferior Temporal Cortex Representations Improves Model-to-Human Behavioral Alignment and Adversarial Robustness
• CVPR 2023Adjustment and Alignment for Unbiased Open Set Domain Adaptation
• NeurIPS 2026$\mathcal{P}$Torch: Narrowing the Gap Between Projection and Gradient-Based Learning
• ICML 2026Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment
• ICLR 2026A2D: Any-Order, Any-Step Safety Alignment for Diffusion Language Models
• CVPR 2026AlignPose Generalizable 6D Pose Estimation via Multi-view Feature-metric Alignment

全部对齐与安全论文

共 1,882 篇，第 26 / 38 页

1251. ICML 2026Reward-free Alignment for Conflicting Objectives
1252. ICML 2026RGGT: A Generative-Prior-Guided Transformer for Unified Rigid and Non-Rigid Point Cloud Registration
1253. ICML 2026Risk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising Utility
1254. ICML 2026RLSF-V: Mitigating Hallucinations in MLLMs via Fuzzy Semantic Self-Feedback
1255. ICML 2026Robust Harmful Features Under Jailbreak Attacks: Mechanistic Evidence from Attention Head Specialization in Large Language Models
1256. ICML 2026Robust Inter-Series Dependency Modeling for Time Series Forecasting via Information-Theoretic Alignment
1257. ICML 2026Robust Linear Dueling Bandits with Post-serving Context under Unknown Delays and Adversarial Corruptions
1258. ICML 2026Romberg-Extrapolated Zeroth-Order Gradient Estimator: Higher-Order Bias Reduction with Preserved Leading Directional Variance
1259. ICML 2026Row-Stochastic Matrices Can Provably Outperform Doubly Stochastic Matrices in Decentralized Learning
1260. ICML 2026RSA-CP: Efficient Conformal Prediction in Small-Sample Regimes via Random Score Alignment
1261. ICML 2026RSPO: Regularized Self-Play Alignment of Large Language Models
1262. ICML 2026SafeHarbor: Defining Precise Decision Boundaries via Hierarchical Memory-Augmented Guardrail for LLM Agent Safety
1263. ICML 2026SafeSci: Safety Evaluation of Large Language Models in Science Domains and Beyond
1264. ICML 2026SafeSearch: Automated Red-Teaming of LLM-Based Search Agents
1265. ICML 2026SafeSeek: Universal Attribution of Safety Circuits in Language Models
1266. ICML 2026SafeSpec: Fast and Safe LLM via Dynamic Reflective Sampling
1267. ICML 2026Safety Alignment of LMs via Non-cooperative Games
1268. ICML 2026Safety Anchor: Defending Harmful Fine-tuning via Geometric Bottlenecks
1269. ICML 2026Safety Game: Inference-Time Alignment of Black-Box LLMs via Constrained Optimization
1270. ICML 2026Safety-Efficacy Trade Off: Robustness against Data-Poisoning
1271. ICML 2026Same Question, Different Lies: Cross-Context Consistency (C³) for Black-Box Sandbagging Detection
1272. ICML 2026SARL: Structure-Aligned Reinforcement Learning for Bridging the Perception-Action Gap in Airspace
1273. ICML 2026SaTeen: Learning Structural Alignment for Continual Test-Time Adaptation
1274. ICML 2026Scalable and Differentiable Point-Cloud Registration Using Maximum Mean Discrepancy
1275. ICML 2026Scalable and Interpretable Representation Alignment with Ordinal Similarity
1276. ICML 2026scDataset: Scalable Data Loading for Deep Learning on Large-Scale Single-Cell Omics
1277. ICML 2026Securing Multimodal AI through Internal Information Decomposition
1278. ICML 2026Segment-driven Structural Induction and Semantic Alignment for Heterogeneous Tabular Representation
1279. ICML 2026Selective Coupling of Decoupled Informative Regions: Masked Attention Alignment for Data-Free Quantization of Vision Transformers
1280. ICML 2026Self-Guidance: Enhancing Neural Codecs via Decoder Manifold Alignment
1281. ICML 2026SEPS: Semantic-Enhanced Patch Slimming Framework for Fine-Grained Cross-Modal Alignment
1282. ICML 2026SHARP-Q: Spectral Hessian Alignment and Rectification for Post-training Quantization
1283. ICML 2026Sign Lock-In: Randomly Initialized Weight Signs Persist and Bottleneck Sub-Bit Model Compression
1284. ICML 2026Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards
1285. ICML 2026SIPO: Stabilized and Improved Preference Optimization for Aligning Diffusion Models
1286. ICML 2026SMILE: Extended Deep Submodular Function-Based Instruction and In-context Learning Demonstration Selection
1287. ICML 2026Solving the Offline and Online Min-Max Problem of Non-smooth Submodular-Concave Functions: A Zeroth-Order Approach
1288. ICML 2026SOTAlign: Semi-Supervised Alignment of Unimodal Vision and Language Models via Optimal Transport
1289. ICML 2026SPARD: Defending Harmful Fine-Tuning Attack via Safety Projection with Relevance–Diversity Data Selection
1290. ICML 2026Sparse Models, Sparse Safety: Unsafe Routes in Mixture-of-Experts LLMs
1291. ICML 2026Sparse Tokens Suffice: Jailbreaking Audio Language Models via Token-Aware Gradient Optimization
1292. ICML 2026Speech-Audio Compositional Attacks on Multimodal LLMs and Their Defense with SALMONN-Guard
1293. ICML 2026Split Personality Training: Revealing Latent Knowledge Through Alternate Personalities
1294. ICML 2026Spurious Correlation Learning in Preference Optimization: Mechanisms, Consequences, and Mitigation via Tie Training
1295. ICML 2026SSA: Sparse Sparse Attention by Aligning Full and Sparse Attention Outputs in Feature Space
1296. ICML 2026Stabilizing PPO via Latent-Space Regularization and KDE-Driven Exploration
1297. ICML 2026Stable Spectral Copula Alignment for Robust Multimodal Learning
1298. ICML 2026Stable Velocity: A Variance Perspective on Flow Matching
1299. ICML 2026Stable-GFlowNet: Toward Diverse and Robust LLM Red-Teaming via Contrastive Trajectory Balance
1300. ICML 2026STARE: Step-wise Temporal Alignment and Red-teaming Engine for Multi-modal Toxicity Attack

常见问题

2023–2026 年四大会议一共有多少篇对齐与安全论文？

共 4,558 篇，按年份分别是 2023 年 438 篇、2024 年 875 篇、2025 年 1,363 篇、2026 年 1,882 篇。

NeurIPS 2026 有多少篇对齐与安全论文？

NeurIPS 2026 有 480 篇对齐与安全论文，上一届是 455 篇。

ICLR 2026 有多少篇对齐与安全论文？

ICLR 2026 有 623 篇对齐与安全论文，上一届是 473 篇。

ICML 2026 有多少篇对齐与安全论文？

ICML 2026 有 551 篇对齐与安全论文，上一届是 300 篇。

CVPR 2026 有多少篇对齐与安全论文？

CVPR 2026 有 228 篇对齐与安全论文，上一届是 135 篇。

哪个会议的对齐与安全论文最多？

NeurIPS，2023–2026 年共 1,421 篇。