风险建模论文趋势 2023–2026
风控模型
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 115 篇被归入该主题的论文。
主要结论
- 风险建模论文从 2023 年的 15 篇增长到 2026 年的 48 篇，三年间扩大到约 3.2 倍。
- NeurIPS 是发风险建模论文最多的会议，2023–2026 年共 47 篇，占四个会议总量的 41%。
- 交叉方向里涨得最快的是强化学习：同时属于两个主题的论文从 2023 年的 4 篇增至 2026 年的 12 篇。
- 在「金融与经济」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的宏观预测是 8.0 倍，风险建模是 3.2 倍。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 7 | 2 | 6 | 0 | 15 |
| 2024 | 6 | 5 | 10 | 5 | 26 |
| 2025 | 9 | 9 | 8 | 0 | 26 |
| 2026 | 25 | 6 | 14 | 3 | 48 |
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 2026CCTimeBoost: Learning Time-Varying Relative Risk with Case-Control Boosted Trees
- ICML 2026A Machine-Learned Comorbidity Index
- ICLR 2026SAFER: Risk-Constrained Sample-then-Filter in Large Language Models
- CVPR 2026Spectral Conformal Risk Control Distribution-Free Tail Guarantees via Bayesian Quadrature
- NeurIPS 2025Generating Informative Samples for Risk-Averse Fine-Tuning of Downstream Tasks
- ICML 2025Compositional Risk Minimization
- ICLR 2025Conformal Generative Modeling with Improved Sample Efficiency through Sequential Greedy Filtering
- NeurIPS 2024Evaluating language models as risk scores
- ICML 2024Differentially Private Worst-group Risk Minimization
- ICLR 2024SILO Language Models: Isolating Legal Risk In a Nonparametric Datastore
- CVPR 2024Self-Calibrating Vicinal Risk Minimisation for Model Calibration
- NeurIPS 2023Optimal Excess Risk Bounds for Empirical Risk Minimization on $p$-Norm Linear Regression
- ICML 2023Fast Excess Risk Rates via Offset Rademacher Complexity
- ICLR 2023Avoiding spurious correlations via logit correction
- NeurIPS 2026Hidden Tails: Certifying Tail-Risk Claims under Selective Labels
- ICML 2026Disentangling Latent Risk Pathways via Bayesian Hypergraph Inference
- ICLR 2026Conformalized Decision Risk Assessment
- CVPR 2026CoRiM Conflict-driven Risk Minimization for Dynamic Multimodal Fusion
- NeurIPS 2025Risk Bounds For Distributional Regression
- ICML 2025Structure-informed Risk Minimization for Robust Ensemble Learning
全部风险建模论文
共 26 篇，第 1 / 1 页
- CVPR 2024PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization
- CVPR 2024Seeing Unseen: Discover Novel Biomedical Concepts via Geometry-Constrained Probabilistic Modeling
- CVPR 2024Self-Calibrating Vicinal Risk Minimisation for Model Calibration
- CVPR 2024Task-Driven Wavelets using Constrained Empirical Risk Minimization
- CVPR 2024XFibrosis: Explicit Vessel-Fiber Modeling for Fibrosis Staging from Liver Pathology Images
- ICLR 2024Conformal Risk Control
- ICLR 2024Meta-VBO: Utilizing Prior Tasks in Optimizing Risk Measures with Gaussian Processes
- ICLR 2024Non-Exchangeable Conformal Risk Control
- ICLR 2024Prompt Risk Control: A Rigorous Framework for Responsible Deployment of Large Language Models
- ICLR 2024SILO Language Models: Isolating Legal Risk In a Nonparametric Datastore
- ICML 2024Bounding the Excess Risk for Linear Models Trained on Marginal-Preserving, Differentially-Private, Synthetic Data
- ICML 2024Differentially Private Worst-group Risk Minimization
- ICML 2024Invariant Risk Minimization Is A Total Variation Model
- ICML 2024On the Asymptotic Distribution of the Minimum Empirical Risk
- ICML 2024Pessimism Meets Risk: Risk-Sensitive Offline Reinforcement Learning
- ICML 2024Position: Enforced Amnesia as a Way to Mitigate the Potential Risk of Silent Suffering in the Conscious AI
- ICML 2024Provably Efficient Partially Observable Risk-sensitive Reinforcement Learning with Hindsight Observation
- ICML 2024Risk Aware Benchmarking of Large Language Models
- ICML 2024Risk-Sensitive Reward-Free Reinforcement Learning with CVaR
- ICML 2024Stereo Risk: A Continuous Modeling Approach to Stereo Matching
- NeurIPS 2024$\epsilon$-Softmax: Approximating One-Hot Vectors for Mitigating Label Noise
- NeurIPS 2024Evaluating language models as risk scores
- NeurIPS 2024Fast yet Safe: Early-Exiting with Risk Control
- NeurIPS 2024Risk-sensitive control as inference with Rényi divergence
- NeurIPS 2024Spectral-Risk Safe Reinforcement Learning with Convergence Guarantees
- NeurIPS 2024Universal Rates of Empirical Risk Minimization
在 Claude、Cursor 里直接问这个趋势
接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。
可以直接复制这句话问 AI
用 paperlist 查风险建模在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？
AI 会调用的工具：get_topic_stats {"topic": "risk-modeling"}
常见问题
- 2023–2026 年四大会议一共有多少篇风险建模论文？
- 共 115 篇，按年份分别是 2023 年 15 篇、2024 年 26 篇、2025 年 26 篇、2026 年 48 篇。
- NeurIPS 2026 有多少篇风险建模论文？
- NeurIPS 2026 有 25 篇风险建模论文，上一届是 9 篇。
- ICLR 2026 有多少篇风险建模论文？
- ICLR 2026 有 6 篇风险建模论文，上一届是 9 篇。
- ICML 2026 有多少篇风险建模论文？
- ICML 2026 有 14 篇风险建模论文，上一届是 8 篇。
- CVPR 2026 有多少篇风险建模论文？
- CVPR 2026 有 3 篇风险建模论文。
- 哪个会议的风险建模论文最多？
- NeurIPS，2023–2026 年共 47 篇。