差分隐私论文趋势 2023–2026

主要结论

• 差分隐私论文从 2023 年的 117 篇降到 2026 年的 103 篇，热度在回落。
• NeurIPS 是发差分隐私论文最多的会议，2023–2026 年共 217 篇，占四个会议总量的 45%。
• 交叉方向里涨得最快的是社会层面 · accountability transparency and interpretability：同时属于两个主题的论文从 2023 年的 0 篇增至 2026 年的 22 篇。
• 在「联邦与隐私」的 5 个主题里，它的 2026 年论文数排第 3；同组增长最快的安全聚合是 1.8 倍，差分隐私是 0.9 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 Unified Framework for Adversary-Aware Differential Privacy Bounds
• ICML 2026Optimal conversion from Rényi Differential Privacy to $f$-Differential Privacy
• ICLR 2026Benchmarking Empirical Privacy Protection for Adaptations of Large Language Models
• CVPR 2026Fully Decentralized Certified Unlearning
• NeurIPS 2025Continual Release Moment Estimation with Differential Privacy
• ICML 2025Leveraging Per-Instance Privacy for Machine Unlearning
• ICLR 2025Data-adaptive Differentially Private Prompt Synthesis for In-Context Learning
• CVPR 2025Focal Split: Untethered Snapshot Depth from Differential Defocus
• NeurIPS 2024$\texttt{pfl-research}$: simulation framework for accelerating research in Private Federated Learning
• ICML 2024Delving into Differentially Private Transformer
• ICLR 2024Decoupled Marked Temporal Point Process using Neural Ordinary Differential Equations
• CVPR 2024Affine Equivariant Networks Based on Differential Invariants
• NeurIPS 2023(Amplified) Banded Matrix Factorization: A unified approach to private training
• ICML 2023Bayesian Estimation of Differential Privacy
• ICLR 2023A Statistical Framework for Personalized Federated Learning and Estimation: Theory, Algorithms, and Privacy
• CVPR 2023Learning To Generate Image Embeddings With User-Level Differential Privacy
• NeurIPS 2026Asymptotically Optimal Best Arm Identification with Fixed-Budget under Differential Privacy
• ICML 2026Optimization, Generalization and Differential Privacy Bounds for Gradient Descent on Kolmogorov–Arnold Networks
• ICLR 2026Beyond Membership: Limitations of Add/Remove Adjacency in Differential Privacy
• CVPR 2026LDP-Slicing Local Differential Privacy for Images via Randomized Bit-Plane Slicing

全部差分隐私论文

共 140 篇，第 1 / 3 页

1. ICML 2026Accuracy-First Rényi Differential Privacy and Post-Processing Immunity
2. ICML 2026Accurate, Private, Secure, Federated U-statistics with Higher Degree
3. ICML 2026Barriers to Counterfactual Credit Attribution for Autoregressive Models
4. ICML 2026Censoring with Plausible Deniability: Asymmetric Local Privacy for Multi-Category CDF Estimation
5. ICML 2026Content-Style Identification via Differential Independence
6. ICML 2026Decoupling Regularization and Privacy in Differentially Private Ridge Regression and ERM
7. ICML 2026Differentially Private Continual Release with Relative Error
8. ICML 2026Differentially Private Cross-Silo Recommendation from Implicit Feedback
9. ICML 2026Differentially Private Geodesic Regression
10. ICML 2026Differentially Private Preference Data Synthesis for Large Language Model Alignment
11. ICML 2026Differentially Private Range Subgraph Counting
12. ICML 2026Differentially Private Submodular Maximization with a Knapsack Constraint
13. ICML 2026Differentially Private Synthetic Data via APIs 4: Tabular Data
14. ICML 2026DP-KFC: Data-Free Preconditioning for Privacy-Preserving Deep Learning
15. ICML 2026Efficient DP-SGD for LLMs with Randomized Clipping
16. ICML 2026Efficient Public Verification of Private ML via Regularization
17. ICML 2026FIBER: A Differentially Private Optimizer with Filter-Aware Innovation Bias Correction
18. ICML 2026Finding Differentially Private Second Order Stationary Points in Stochastic Minimax Optimization
19. ICML 2026Keeping a Secret Requires a Good Memory: Unconditional Streaming Lower-Bounds for Differentially Private Algorithms
20. ICML 2026LAPRAS : Learning-Augmented PRivate Answering for linear query Streams.
21. ICML 2026Mind the Gap: Mixtures of Gaussians in Approximate Differential Privacy
22. ICML 2026Near-Optimal Private Linear Regression via Iterative Hessian Mixing
23. ICML 2026OptiFluence: Principled Design of Privacy Canaries
24. ICML 2026Optimal conversion from Rényi Differential Privacy to $f$-Differential Privacy
25. ICML 2026Optimal Domain-Aware Privacy Mechanisms for Synthetic Data Generation
26. ICML 2026Optimization, Generalization and Differential Privacy Bounds for Gradient Descent on Kolmogorov–Arnold Networks
27. ICML 2026Persuasive Privacy
28. ICML 2026PISA: Privacy-Preserving Split Adaptation with Model IP Protection
29. ICML 2026PRISM: Gauge-Invariant Tangent-Space Differentially Private LoRA
30. ICML 2026Privacy Amplification in Differentially Private Zeroth-Order Optimization with Hidden States
31. ICML 2026Privacy-Aware Data Integration for Enhanced Quantile Inference under Heterogeneity
32. ICML 2026Private and Stable Test-time Adaptation with Differential Privacy
33. ICML 2026Private Learning with Public Feature Conditioning
34. ICML 2026Privately Fine-Tuned LLMs Preserve Temporal Dynamics in Tabular Data
35. ICML 2026PrivCode++: Latent-Conditioned Differentially Private Code Generation for Comprehensive Guarantees
36. ICML 2026Regret-Based Federated Causal Discovery with Unknown Interventions
37. ICML 2026Revisiting ML Training under Fully Homomorphic Encryption: Convergence Guarantees, Differential Privacy, and Efficient Algorithms
38. ICML 2026scDEBART: Predicting in silico Single-Cell Perturbation Responses via Large-Scale Differential Expression Learning
39. ICML 2026Shuffling-Aware Optimization for Private Vector Mean Estimation
40. ICML 2026SlaClip: Gradient Norm Slacks can be Indicator for Adaptive Clipping in DP-SGD
41. ICML 2026Solving Positive Linear Programs with Differential Privacy
42. ICML 2026Understanding Private Learning From Feature Perspective
43. ICML 2025Adapting to Linear Separable Subsets with Large-Margin in Differentially Private Learning
44. ICML 2025Approximate Differential Privacy of the $\ell_2$ Mechanism
45. ICML 2025Auditing $f$-differential privacy in one run
46. ICML 2025Breaking the $n^{1.5}$ Additive Error Barrier for Private and Efficient Graph Sparsification via Private Expander Decomposition
47. ICML 2025Breaking the Quadratic Barrier: Robust Cardinality Sketches for Adaptive Queries
48. ICML 2025Cape: Context-Aware Prompt Perturbation Mechanism with Differential Privacy
49. ICML 2025Certification for Differentially Private Prediction in Gradient-Based Training
50. ICML 2025Clients Collaborate: Flexible Differentially Private Federated Learning with Guaranteed Improvement of Utility-Privacy Trade-off

常见问题

2023–2026 年四大会议一共有多少篇差分隐私论文？

共 479 篇，按年份分别是 2023 年 117 篇、2024 年 113 篇、2025 年 146 篇、2026 年 103 篇。

NeurIPS 2026 有多少篇差分隐私论文？

NeurIPS 2026 有 23 篇差分隐私论文，上一届是 66 篇。

ICLR 2026 有多少篇差分隐私论文？

ICLR 2026 有 31 篇差分隐私论文，上一届是 29 篇。

ICML 2026 有多少篇差分隐私论文？

ICML 2026 有 42 篇差分隐私论文，上一届是 47 篇。

CVPR 2026 有多少篇差分隐私论文？

CVPR 2026 有 7 篇差分隐私论文，上一届是 4 篇。

哪个会议的差分隐私论文最多？

NeurIPS，2023–2026 年共 217 篇。