社会层面 · fairness论文趋势 2023–2026
AI 社会议题：fairness
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 358 篇被归入该主题的论文。
主要结论
- 社会层面 · fairness论文从 2023 年的 36 篇增长到 2026 年的 111 篇，三年间扩大到约 3.1 倍。
- NeurIPS 是发社会层面 · fairness论文最多的会议，2023–2026 年共 180 篇，占四个会议总量的 50%。
- 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 1 篇增至 2026 年的 56 篇。
- 在「对齐与负责任 AI」的 5 个主题里，它的 2026 年论文数排第 5；同组增长最快的社会层面 · safety是 93 倍，社会层面 · fairness是 3.1 倍。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 19 | 8 | 7 | 2 | 36 |
| 2024 | 56 | 9 | 11 | 3 | 79 |
| 2025 | 62 | 8 | 61 | 1 | 132 |
| 2026 | 43 | 4 | 60 | 4 | 111 |
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 2026Beyond Feature Disruption: Boundary-Diverting Unlearnable Examples against Linear Probing
- ICML 2026A Game-Theoretic Framework for Measuring and Explaining Metric Compatibility in Fair Machine Learning
- ICLR 2026Bi-directional Bias Attribution: Debiasing Large Language Models without Modifying Prompts
- CVPR 2026Omni-MMSI Toward Identity-attributed Social Interaction Understanding
- NeurIPS 2025Discretization-free Multicalibration through Loss Minimization over Tree Ensembles
- ICML 2025Dissecting Submission Limit in Desk-Rejections: A Mathematical Analysis of Fairness in AI Conference Policies
- ICLR 2025Prompting Fairness: Integrating Causality to Debias Large Language Models
- CVPR 2025Revisiting Fairness in Multitask Learning: A Performance-Driven Approach for Variance Reduction
- NeurIPS 2024A Closer Look at AUROC and AUPRC under Class Imbalance
- ICML 2024Position: Scarce Resource Allocations That Rely On Machine Learning Should Be Randomized
- ICLR 2024Procedural Fairness Through Decoupling Objectionable Data Generating Components
- CVPR 2024Preserving Fairness Generalization in Deepfake Detection
- NeurIPS 2023[Re] Fairness Guarantees under Demographic Shift
- ICML 2023Fairness in Streaming Submodular Maximization over a Matroid Constraint
- ICLR 2023Fairness-aware Contrastive Learning with Partially Annotated Sensitive Attributes
- CVPR 2023Improving Fairness in Facial Albedo Estimation via Visual-Textual Cues
- NeurIPS 2026Chatter Attack: Resource Consumption Attack for Large Language Models
- ICML 2026Adaptive Generation of Bias-Eliciting Questions for LLMs
- ICLR 2026Bi-Criteria Metric Distortion
- CVPR 2026Decoupling Bias Aligning Distributions Synergistic Fairness Optimization for Deepfake Detection
全部社会层面 · fairness论文
共 61 篇，第 2 / 2 页
- ICML 2025The Disparate Benefits of Deep Ensembles
- ICML 2025The Elicitation Game: Evaluating Capability Elicitation Techniques
- ICML 2025The Illusion of Role Separation: Hidden Shortcuts in LLM Role Learning (and How to Fix Them)
- ICML 2025The Value of Prediction in Identifying the Worst-Off
- ICML 2025TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference
- ICML 2025Towards Attributions of Input Variables in a Coalition
- ICML 2025Two Tickets are Better than One: Fair and Accurate Hiring Under Strategic LLM Manipulations
- ICML 2025Uncertainty-Based Extensible Codebook for Discrete Federated Learning in Heterogeneous Data Silos
- ICML 2025What makes an Ensemble (Un) Interpretable?
- ICML 2025X-Hacking: The Threat of Misguided AutoML
- ICML 2025You Get What You Give: Reciprocally Fair Federated Learning
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 查社会层面 · fairness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": "social-fairness"}
常见问题
- 2023–2026 年四大会议一共有多少篇社会层面 · fairness论文？
- 共 358 篇，按年份分别是 2023 年 36 篇、2024 年 79 篇、2025 年 132 篇、2026 年 111 篇。
- NeurIPS 2026 有多少篇社会层面 · fairness论文？
- NeurIPS 2026 有 43 篇社会层面 · fairness论文，上一届是 62 篇。
- ICLR 2026 有多少篇社会层面 · fairness论文？
- ICLR 2026 有 4 篇社会层面 · fairness论文，上一届是 8 篇。
- ICML 2026 有多少篇社会层面 · fairness论文？
- ICML 2026 有 60 篇社会层面 · fairness论文，上一届是 61 篇。
- CVPR 2026 有多少篇社会层面 · fairness论文？
- CVPR 2026 有 4 篇社会层面 · fairness论文，上一届是 1 篇。
- 哪个会议的社会层面 · fairness论文最多？
- NeurIPS，2023–2026 年共 180 篇。