特征归因论文趋势 2023–2026

主要结论

• 特征归因论文从 2023 年的 118 篇增长到 2026 年的 219 篇，三年间扩大到约 1.9 倍。
• NeurIPS 是发特征归因论文最多的会议，2023–2026 年共 193 篇，占四个会议总量的 28%。
• 交叉方向里涨得最快的是理论 · learning theory：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 14 篇。
• 在「可解释性」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的机制解释是 8.1 倍，特征归因是 1.9 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 2026ADAPT: Hybrid Prompt Optimization for LLM Feature Visualization
• ICML 2026Activation with Intrinsic-Extrinsic Consensus
• ICLR 2026AnyUp: Universal Feature Upsampling
• CVPR 2026A Training-Free Style-Personalization via SVD-Based Feature Decomposition
• NeurIPS 2025Alternating Gradient Flows: A Theory of Feature Learning in Two-layer Neural Networks
• ICML 2025BCE vs. CE in Deep Feature Learning
• ICLR 2025A Differentiable Rank-Based Objective for Better Feature Learning
• CVPR 2025A Simple Data Augmentation for Feature Distribution Skewed Federated Learning
• NeurIPS 2024$\texttt{dattri}$: A Library for Efficient Data Attribution
• ICML 2024Asymptotics of feature learning in two-layer networks after one gradient-step
• ICLR 2024An Intuitive Multi-Frequency Feature Representation for SO(3)-Equivariant Networks
• CVPR 2024DYSON: Dynamic Feature Space Self-Organization for Online Task-Free Class Incremental Learning
• NeurIPS 2023$\mathcal{M}^4$: A Unified XAI Benchmark for Faithfulness Evaluation of Feature Attribution Methods across Metrics, Modalities and Models
• ICML 2023Decoding Layer Saliency in Language Transformers
• ICLR 2023Can We Faithfully Represent Absence States to Compute Shapley Values on a DNN?
• CVPR 2023A Practical Upper Bound for the Worst-Case Attribution Deviations
• NeurIPS 2026Adaptive Feature Propagation for Attribute-Missing Graph Clustering
• ICML 2026Aggregate Models, Not Explanations: Improving Feature Importance Estimation
• ICLR 2026Attribution-Guided Decoding
• CVPR 2026FBTA Enabling Single-GPU End-to-End Gigapixel WSI Classification with Feature Bridging

全部特征归因论文

共 118 篇，第 2 / 3 页

51. ICLR 2023Feature selection and low test error in shallow low-rotation ReLU networks
52. ICLR 2023FedFA: Federated Feature Augmentation
53. ICLR 2023Fooling SHAP with Stealthily Biased Sampling
54. ICLR 2023Function-Consistent Feature Distillation
55. ICLR 2023GEASS: Neural causal feature selection for high-dimensional biological data
56. ICLR 2023Inequality phenomenon in $l_{\infty}$-adversarial training, and its unrealized threats
57. ICLR 2023Interpretability with full complexity by constraining feature information
58. ICLR 2023Learning a Data-Driven Policy Network for Pre-Training Automated Feature Engineering
59. ICLR 2023Learning Multimodal Data Augmentation in Feature Space
60. ICLR 2023Long-Tailed Learning Requires Feature Learning
61. ICLR 2023MaskFusion: Feature Augmentation for Click-Through Rate Prediction via Input-adaptive Mask Fusion
62. ICLR 2023Neural Networks Efficiently Learn Low-Dimensional Representations with SGD
63. ICLR 2023One-Pixel Shortcut: On the Learning Preference of Deep Neural Networks
64. ICLR 2023Over-Training with Mixup May Hurt Generalization
65. ICLR 2023PandA: Unsupervised Learning of Parts and Appearances in the Feature Maps of GANs
66. ICLR 2023Personalized Federated Learning with Feature Alignment and Classifier Collaboration
67. ICLR 2023Predictive Inference with Feature Conformal Prediction
68. ICLR 2023Re-calibrating Feature Attributions for Model Interpretation
69. ICLR 2023Sequential Attention for Feature Selection
70. ICLR 2023Temporal Dependencies in Feature Importance for Time Series Prediction
71. ICLR 2023The Curious Case of Benign Memorization
72. ICLR 2023The Onset of Variance-Limited Behavior for Networks in the Lazy and Rich Regimes
73. ICML 2023Decoding Layer Saliency in Language Transformers
74. ICML 2023Defects of Convolutional Decoder Networks in Frequency Representation
75. ICML 2023DRCFS: Doubly Robust Causal Feature Selection
76. ICML 2023Faith-Shap: The Faithful Shapley Interaction Index
77. ICML 2023Feature Expansion for Graph Neural Networks
78. ICML 2023Feature learning in deep classifiers through Intermediate Neural Collapse
79. ICML 2023Feature Programming for Multivariate Time Series Prediction
80. ICML 2023Global optimality of Elman-type RNNs in the mean-field regime
81. ICML 2023GRAFENNE: Learning on Graphs with Heterogeneous and Dynamic Feature Sets
82. ICML 2023Graph Positional Encoding via Random Feature Propagation
83. ICML 2023Hybrid Energy Based Model in the Feature Space for Out-of-Distribution Detection
84. ICML 2023Learning Perturbations to Explain Time Series Predictions
85. ICML 2023Learning to Maximize Mutual Information for Dynamic Feature Selection
86. ICML 2023LESS-VFL: Communication-Efficient Feature Selection for Vertical Federated Learning
87. ICML 2023MetaModulation: Learning Variational Feature Hierarchies for Few-Shot Learning with Fewer Tasks
88. ICML 2023Monge, Bregman and Occam: Interpretable Optimal Transport in High-Dimensions with Feature-Sparse Maps
89. ICML 2023Simple and Fast Group Robustness by Automatic Feature Reweighting
90. ICML 2023Sketched Ridgeless Linear Regression: The Role of Downsampling
91. ICML 2023SRATTA: Sample Re-ATTribution Attack of Secure Aggregation in Federated Learning.
92. ICML 2023The Benefits of Mixup for Feature Learning
93. ICML 2023Towards credible visual model interpretation with path attribution
94. NeurIPS 2023$\mathcal{M}^4$: A Unified XAI Benchmark for Faithfulness Evaluation of Feature Attribution Methods across Metrics, Modalities and Models
95. NeurIPS 2023A Neural Collapse Perspective on Feature Evolution in Graph Neural Networks
96. NeurIPS 2023A Performance-Driven Benchmark for Feature Selection in Tabular Deep Learning
97. NeurIPS 2023Adversarial Learning for Feature Shift Detection and Correction
98. NeurIPS 2023Better Private Linear Regression Through Better Private Feature Selection
99. NeurIPS 2023Discriminative Feature Attributions: Bridging Post Hoc Explainability and Inherent Interpretability
100. NeurIPS 2023Error Bounds for Learning with Vector-Valued Random Features

常见问题

2023–2026 年四大会议一共有多少篇特征归因论文？

共 695 篇，按年份分别是 2023 年 118 篇、2024 年 168 篇、2025 年 190 篇、2026 年 219 篇。

NeurIPS 2026 有多少篇特征归因论文？

NeurIPS 2026 有 61 篇特征归因论文，上一届是 57 篇。

ICLR 2026 有多少篇特征归因论文？

ICLR 2026 有 53 篇特征归因论文，上一届是 42 篇。

ICML 2026 有多少篇特征归因论文？

ICML 2026 有 56 篇特征归因论文，上一届是 51 篇。

CVPR 2026 有多少篇特征归因论文？

CVPR 2026 有 49 篇特征归因论文，上一届是 40 篇。

哪个会议的特征归因论文最多？

NeurIPS，2023–2026 年共 193 篇。