概念解释论文趋势 2023–2026

主要结论

• 概念解释论文从 2023 年的 91 篇增长到 2026 年的 117 篇，三年间扩大到约 1.3 倍。
• NeurIPS 是发概念解释论文最多的会议，2023–2026 年共 134 篇，占四个会议总量的 39%。
• 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 4 篇增至 2026 年的 23 篇。
• 在「可解释性」的 5 个主题里，它的 2026 年论文数排第 3；同组增长最快的机制解释是 8.1 倍，概念解释是 1.3 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 2026CENDRe: Concept Extraction with Natural Domain Representations
• ICML 2026Evidential Copula Concept Embedding Models
• ICLR 2026Decoupling the Class Label and the Target Concept in Machine Unlearning
• CVPR 2026Agile Deliberation Concept Deliberation for Subjective Visual Classification
• NeurIPS 2025An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations
• ICML 2025Concept-Centric Token Interpretation for Vector-Quantized Generative Models
• ICLR 2025Causal Concept Graph Models: Beyond Causal Opacity in Deep Learning
• CVPR 2025Discovering Fine-Grained Visual-Concept Relations by Disentangled Optimal Transport Concept Bottleneck Models
• NeurIPS 2024[Re] On the Reproducibility of Post-Hoc Concept Bottleneck Models
• ICML 2024ContPhy: Continuum Physical Concept Learning and Reasoning from Videos
• ICLR 2024AttEXplore: Attribution for Explanation with model parameters eXploration
• CVPR 2024Fine-Grained Bipartite Concept Factorization for Clustering
• NeurIPS 2023[Re] Hierarchical Shrinkage: Improving the Accuracy and Interpretability of Tree-Based Methods
• ICML 2023A Closer Look at the Intervention Procedure of Concept Bottleneck Models
• ICLR 2023Concept Gradient: Concept-based Interpretation Without Linear Assumption
• CVPR 2023CRAFT: Concept Recursive Activation FacTorization for Explainability
• NeurIPS 2026Concept frustration: Aligning human concepts and machine representations
• ICML 2026On the Variability of Concept Activation Vectors
• ICLR 2026Hierarchical Concept-based Interpretable Models
• CVPR 2026DCoAR Deep Concept Injection into Unified Autoregressive Models for Personalized

全部概念解释论文

共 134 篇，第 2 / 3 页

51. NeurIPS 2025Disentangling Superpositions: Interpretable Brain Encoding Model with Sparse Concept Atoms
52. NeurIPS 2025EraseFlow: Learning Concept Erasure Policies via GFlowNet-Driven Alignment
53. NeurIPS 2025FACE: Faithful Automatic Concept Extraction
54. NeurIPS 2025FaCT: Faithful Concept Traces for Explaining Neural Network Decisions
55. NeurIPS 2025Generalization vs Specialization under Concept Shift
56. NeurIPS 2025Learning to Condition: A Neural Heuristic for Scalable MPE Inference
57. NeurIPS 2025Object-Centric Concept-Bottlenecks
58. NeurIPS 2025Obliviator Reveals the Cost of Nonlinear Guardedness in Concept Erasure
59. NeurIPS 2025Preserving Task-Relevant Information Under Linear Concept Removal
60. NeurIPS 2025Projecting Assumptions: The Duality Between Sparse Autoencoders and Concept Geometry
61. NeurIPS 2025RCCDA: Adaptive Model Updates in the Presence of Concept Drift under a Constrained Resource Budget
62. NeurIPS 2025Shortcuts and Identifiability in Concept-based Models from a Neuro-Symbolic Lens
63. NeurIPS 2025Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space
64. NeurIPS 2025When Are Concepts Erased From Diffusion Models?
65. NeurIPS 2025Why Knowledge Distillation Works in Generative Models: A Minimal Working Explanation
66. NeurIPS 2024[Re] On the Reproducibility of Post-Hoc Concept Bottleneck Models
67. NeurIPS 2024A Concept-Based Explainability Framework for Large Multimodal Models
68. NeurIPS 2024A Neural Network Approach for Efficiently Answering Most Probable Explanation Queries in Probabilistic Models
69. NeurIPS 2024A theoretical design of concept sets: improving the predictability of concept bottleneck models
70. NeurIPS 2024Auditing Local Explanations is Hard
71. NeurIPS 2024Beyond Concept Bottleneck Models: How to Make Black Boxes Intervenable?
72. NeurIPS 2024Coarse-to-Fine Concept Bottleneck Models
73. NeurIPS 2024Defensive Unlearning with Adversarial Training for Robust Concept Erasure in Diffusion Models
74. NeurIPS 2024Do LLMs dream of elephants (when told not to)? Latent concept association and associative memory in transformers
75. NeurIPS 2024From Causal to Concept-Based Representation Learning
76. NeurIPS 2024Improving Decision Sparsity
77. NeurIPS 2024Interpretable Concept Bottlenecks to Align Reinforcement Learning Agents
78. NeurIPS 2024Interpretable Concept-Based Memory Reasoning
79. NeurIPS 2024Interpreting CLIP with Sparse Linear Concept Embeddings (SpLiCE)
80. NeurIPS 2024Neural Concept Binder
81. NeurIPS 2024PaCE: Parsimonious Concept Engineering for Large Language Models
82. NeurIPS 2024Relational Concept Bottleneck Models
83. NeurIPS 2024Stochastic Concept Bottleneck Models
84. NeurIPS 2024Towards Multi-dimensional Explanation Alignment for Medical Classification
85. NeurIPS 2024Training for Stable Explanation for Free
86. NeurIPS 2024Uncovering Safety Risks of Large Language Models through Concept Activation Vector
87. NeurIPS 2023[Re] Hierarchical Shrinkage: Improving the Accuracy and Interpretability of Tree-Based Methods
88. NeurIPS 2023$k$-Means Clustering with Distance-Based Privacy
89. NeurIPS 2023A Dynamical System View of Langevin-Based Non-Convex Sampling
90. NeurIPS 2023A Holistic Approach to Unifying Automatic Concept Extraction and Concept Importance Estimation
91. NeurIPS 2023A Theory of Transfer-Based Black-Box Attacks: Explanation and Implications
92. NeurIPS 2023A*Net: A Scalable Path-based Reasoning Approach for Knowledge Graphs
93. NeurIPS 2023Auxiliary Losses for Learning Generalizable Concept-based Models
94. NeurIPS 2023Concept Algebra for (Score-Based) Text-Controlled Generative Models
95. NeurIPS 2023Concept Distillation: Leveraging Human-Centered Explanations for Model Improvement
96. NeurIPS 2023Content-based Unrestricted Adversarial Attack
97. NeurIPS 2023D-Separation for Causal Self-Explanation
98. NeurIPS 2023Decompose Novel into Known: Part Concept Learning For 3D Novel Class Discovery
99. NeurIPS 2023Discriminative Calibration: Check Bayesian Computation from Simulations and Flexible Classifier
100. NeurIPS 2023EMBERSim: A Large-Scale Databank for Boosting Similarity Search in Malware Analysis

常见问题

2023–2026 年四大会议一共有多少篇概念解释论文？

共 342 篇，按年份分别是 2023 年 91 篇、2024 年 63 篇、2025 年 71 篇、2026 年 117 篇。

NeurIPS 2026 有多少篇概念解释论文？

NeurIPS 2026 有 40 篇概念解释论文，上一届是 25 篇。

ICLR 2026 有多少篇概念解释论文？

ICLR 2026 有 24 篇概念解释论文，上一届是 19 篇。

ICML 2026 有多少篇概念解释论文？

ICML 2026 有 20 篇概念解释论文，上一届是 11 篇。

CVPR 2026 有多少篇概念解释论文？

CVPR 2026 有 33 篇概念解释论文，上一届是 16 篇。

哪个会议的概念解释论文最多？

NeurIPS，2023–2026 年共 134 篇。