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 175 篇，第 3 / 4 页

101. ICML 2025TAROT: Targeted Data Selection via Optimal Transport
102. ICML 2025The Complexity of Learning Sparse Superposed Features with Feedback
103. ICML 2025The Importance of Being Lazy: Scaling Limits of Continual Learning
104. ICML 2025Theoretical Limitations of Ensembles in the Age of Overparameterization
105. ICML 2025Token Signature: Predicting Chain-of-Thought Gains with Token Decoding Feature in Large Language Models
106. ICML 2025Towards an Explainable Comparison and Alignment of Feature Embeddings
107. ICML 2025Weak-to-Strong Generalization Even in Random Feature Networks, Provably
108. ICML 2024A New Computationally Efficient Algorithm to solve Feature Selection for Functional Data Classification in High-dimensional Spaces
109. ICML 2024A Theory of Non-Linear Feature Learning with One Gradient Step in Two-Layer Neural Networks
110. ICML 2024Acquisition Conditioned Oracle for Nongreedy Active Feature Acquisition
111. ICML 2024Asymptotics of feature learning in two-layer networks after one gradient-step
112. ICML 2024Asymptotics of Learning with Deep Structured (Random) Features
113. ICML 2024Balancing Feature Similarity and Label Variability for Optimal Size-Aware One-shot Subset Selection
114. ICML 2024Benchmarking Deletion Metrics with the Principled Explanations
115. ICML 2024Catapults in SGD: spikes in the training loss and their impact on generalization through feature learning
116. ICML 2024Complexity Matters: Feature Learning in the Presence of Spurious Correlations
117. ICML 2024Contextual Feature Selection with Conditional Stochastic Gates
118. ICML 2024Critical feature learning in deep neural networks
119. ICML 2024Critical windows: non-asymptotic theory for feature emergence in diffusion models
120. ICML 2024Decoupling Feature Extraction and Classification Layers for Calibrated Neural Networks
121. ICML 2024DFD: Distilling the Feature Disparity Differently for Detectors
122. ICML 2024Don't trust your eyes: on the (un)reliability of feature visualizations
123. ICML 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty
124. ICML 2024Feature Attribution with Necessity and Sufficiency via Dual-stage Perturbation Test for Causal Explanation
125. ICML 2024Feature Contamination: Neural Networks Learn Uncorrelated Features and Fail to Generalize
126. ICML 2024Feature Distribution on Graph Topology Mediates the Effect of Graph Convolution: Homophily Perspective
127. ICML 2024Feature Importance Disparities for Data Bias Investigations
128. ICML 2024Feature Reuse and Scaling: Understanding Transfer Learning with Protein Language Models
129. ICML 2024Helpful or Harmful Data? Fine-tuning-free Shapley Attribution for Explaining Language Model Predictions
130. ICML 2024High-dimensional Linear Bandits with Knapsacks
131. ICML 2024InterpreTabNet: Distilling Predictive Signals from Tabular Data by Salient Feature Interpretation
132. ICML 2024Iterative Search Attribution for Deep Neural Networks
133. ICML 2024Learning 1-Bit Tiny Object Detector with Discriminative Feature Refinement
134. ICML 2024Learning in Feature Spaces via Coupled Covariances: Asymmetric Kernel SVD and Nyström method
135. ICML 2024Local Feature Selection without Label or Feature Leakage for Interpretable Machine Learning Predictions
136. ICML 2024Manifold Integrated Gradients: Riemannian Geometry for Feature Attribution
137. ICML 2024MFTN: A Multi-scale Feature Transfer Network Based on IMatchFormer for Hyperspectral Image Super-Resolution
138. ICML 2024Multi-layer Rehearsal Feature Augmentation for Class-Incremental Learning
139. ICML 2024Multiply-Robust Causal Change Attribution
140. ICML 2024Observable Propagation: Uncovering Feature Vectors in Transformers
141. ICML 2024On dimensionality of feature vectors in MPNNs
142. ICML 2024On the Tractability of SHAP Explanations under Markovian Distributions
143. ICML 2024Pragmatic Feature Preferences: Learning Reward-Relevant Preferences from Human Input
144. ICML 2024Prospector Heads: Generalized Feature Attribution for Large Models & Data
145. ICML 2024ProtoGate: Prototype-based Neural Networks with Global-to-local Feature Selection for Tabular Biomedical Data
146. ICML 2024Provable Benefits of Local Steps in Heterogeneous Federated Learning for Neural Networks: A Feature Learning Perspective
147. ICML 2024Provably Better Explanations with Optimized Aggregation of Feature Attributions
148. ICML 2024Provably Neural Active Learning Succeeds via Prioritizing Perplexing Samples
149. ICML 2024Saliency strikes back: How filtering out high frequencies improves white-box explanations
150. ICML 2024SF-DQN: Provable Knowledge Transfer using Successor Feature for Deep Reinforcement Learning

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。