点击率预测论文趋势 2023–2026
CTR 模型
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 940 篇被归入该主题的论文。
主要结论
- 点击率预测论文从 2023 年的 124 篇增长到 2026 年的 319 篇，三年间扩大到约 2.6 倍。
- NeurIPS 是发点击率预测论文最多的会议，2023–2026 年共 344 篇，占四个会议总量的 37%。
- 交叉方向里涨得最快的是多模态大模型：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 26 篇。
- 在「推荐与广告」的 5 个主题里，它的 2026 年论文数排第 2；同组增长最快的因果推荐是 2.3 倍，点击率预测是 2.6 倍。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 36 | 20 | 32 | 36 | 124 |
| 2024 | 85 | 33 | 77 | 31 | 226 |
| 2025 | 105 | 57 | 65 | 44 | 271 |
| 2026 | 118 | 56 | 77 | 68 | 319 |
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 Comprehensive View of Fairness through Distributional Stability
- ICML 2026A Penalty Approach For Differentiation Through Black-box Quadratic Programming Solvers
- ICLR 2026Adjusting Prediction Model Through Wasserstein Geodesic for Causal Inference
- CVPR 2026DepthFocus Controllable Depth Estimation for See-Through Scenes
- NeurIPS 2025Accident Anticipation via Temporal Occurrence Prediction
- ICML 2025Conformal Prediction with Cellwise Outliers: A Detect-then-Impute Approach
- ICLR 2025Agree to Disagree: Demystifying Homogeneous Deep Ensembles through Distributional Equivalence
- CVPR 2025Adapting to Observation Length of Trajectory Prediction via Contrastive Learning
- NeurIPS 2024[Re] Reproducibility Study of “Explaining Temporal Graph Models Through an Explorer-Navigator Framework"
- ICML 2024Auditing Private Prediction
- ICLR 2024Domain constraints improve risk prediction when outcome data is missing
- CVPR 2024Enhancing Post-training Quantization Calibration through Contrastive Learning
- NeurIPS 2023Aggregating Capacity in FL through Successive Layer Training for Computationally-Constrained Devices
- ICML 2023Conformal Prediction for Federated Uncertainty Quantification Under Label Shift
- ICLR 2023Agree to Disagree: Diversity through Disagreement for Better Transferability
- CVPR 2023Computational Flash Photography Through Intrinsics
- NeurIPS 2026A Kernel Nonconformity Score for Multivariate Conformal Prediction
- ICML 2026Addressing Instrument-Outcome Confounding in Mendelian Randomization through Representation Learning
- ICLR 2026Beyond Uniformity: Regularizing Implicit Neural Representations through a Lipschitz Lens
- CVPR 2026Detect Anything via Next Point Prediction
全部点击率预测论文
共 271 篇，第 6 / 6 页
- NeurIPS 2025Split conformal classification with unsupervised calibration
- NeurIPS 2025SPMDM: Enhancing Masked Diffusion Models through Simplifing Sampling Path
- NeurIPS 2025STAIR: Addressing Stage Misalignment through Temporal-Aligned Preference Reinforcement Learning
- NeurIPS 2025Statistical Inference under Performativity
- NeurIPS 2025Stochastic Forward-Forward Learning through Representational Dimensionality Compression
- NeurIPS 2025Stochastically Dominant Peer Prediction
- NeurIPS 2025SymMaP: Improving Computational Efficiency in Linear Solvers through Symbolic Preconditioning
- NeurIPS 2025TAMI: Taming Heterogeneity in Temporal Interactions for Temporal Graph Link Prediction
- NeurIPS 2025Teaching Transformers to Solve Combinatorial Problems through Efficient Trial & Error
- NeurIPS 2025The $\varphi$ Curve: The Shape of Generalization through the Lens of Norm-based Capacity Control
- NeurIPS 2025Through the Lens: Benchmarking Deepfake Detectors Against Moiré-Induced Distortions
- NeurIPS 2025Timely Clinical Diagnosis through Active Test Selection
- NeurIPS 2025Topology-Aware Conformal Prediction for Stream Networks
- NeurIPS 2025Uncertainty-Calibrated Prediction of Randomly-Timed Biomarker Trajectories with Conformal Bands
- NeurIPS 2025Understanding Fairness and Prediction Error through Subspace Decomposition and Influence Analysis
- NeurIPS 2025Uni-Instruct: One-step Diffusion Model through Unified Diffusion Divergence Instruction
- NeurIPS 2025Universal Sequence Preconditioning
- NeurIPS 2025Unveiling m-Sharpness Through the Structure of Stochastic Gradient Noise
- NeurIPS 2025Valid Selection among Conformal Sets
- NeurIPS 2025VisualLens: Personalization through Task-Agnostic Visual History
- NeurIPS 2025What Does It Take to Build a Performant Selective Classifier?
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 查点击率预测在 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": "click-through-prediction"}
常见问题
- 2023–2026 年四大会议一共有多少篇点击率预测论文？
- 共 940 篇，按年份分别是 2023 年 124 篇、2024 年 226 篇、2025 年 271 篇、2026 年 319 篇。
- NeurIPS 2026 有多少篇点击率预测论文？
- NeurIPS 2026 有 118 篇点击率预测论文，上一届是 105 篇。
- ICLR 2026 有多少篇点击率预测论文？
- ICLR 2026 有 56 篇点击率预测论文，上一届是 57 篇。
- ICML 2026 有多少篇点击率预测论文？
- ICML 2026 有 77 篇点击率预测论文，上一届是 65 篇。
- CVPR 2026 有多少篇点击率预测论文？
- CVPR 2026 有 68 篇点击率预测论文，上一届是 44 篇。
- 哪个会议的点击率预测论文最多？
- NeurIPS，2023–2026 年共 344 篇。