通用机器学习 · generalization论文趋势 2023–2026

主要结论

• 通用机器学习 · generalization论文从 2023 年的 136 篇增长到 2026 年的 152 篇，三年间扩大到约 1.1 倍。
• NeurIPS 是发通用机器学习 · generalization论文最多的会议，2023–2026 年共 207 篇，占四个会议总量的 35%。
• 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 3 篇增至 2026 年的 20 篇。
• 在「通用机器学习」的 6 个主题里，它的 2026 年论文数排第 5；同组增长最快的通用机器学习 · evaluation是 4.3 倍，通用机器学习 · generalization是 1.1 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 2026Asymmetric Generalization in Deep CTR Models: A Block-wise Diagnosis
• ICML 2026Anti-causal domain generalization: Leveraging unlabeled data
• ICLR 2026Generalization Below the Edge of Stability: The Role of Data Geometry
• CVPR 2026PriVi Towards a General-Purpose Video Model for Primate Behavior in
• NeurIPS 2025Automatic Visual Instrumental Variable Learning for Confounding-Resistant Domain Generalization
• ICML 2025Causality Inspired Federated Learning for OOD Generalization
• ICLR 2025Federated Domain Generalization with Data-free On-server Matching Gradient
• CVPR 2025Balanced Direction from Multifarious Choices: Arithmetic Meta-Learning for Domain Generalization
• NeurIPS 2024Benchmarking Out-of-Distribution Generalization Capabilities of DNN-based Encoding Models for the Ventral Visual Cortex.
• ICML 2024Algorithmic Stability Unleashed: Generalization Bounds with Unbounded Losses
• ICLR 2024Adaptive Sharpness-Aware Pruning for Robust Sparse Networks
• CVPR 2024Disentangled Prompt Representation for Domain Generalization
• NeurIPS 2023A unified framework for information-theoretic generalization bounds
• ICML 2023Algorithmic Stability of Heavy-Tailed SGD with General Loss Functions
• ICLR 2023ArCL: Enhancing Contrastive Learning with Augmentation-Robust Representations
• CVPR 2023AltFreezing for More General Video Face Forgery Detection
• NeurIPS 2026Beyond Average Flatness: Domain-wise Flatness for Domain Generalization
• ICML 2026Are Object-Centric Representations Better at Compositional Generalization?
• ICLR 2026Temporal Generalization: A Reality Check
• CVPR 2026The Golden Subspace Where Efficiency Meets Generalization in Continual Test-Time

全部通用机器学习 · generalization论文

共 207 篇，第 2 / 5 页

51. NeurIPS 2026Tokenizer Choice Shapes Generalization in State-Centric Learning for Planning
52. NeurIPS 2026Towards Better Generalization in Lifelong Person Re-Identification with Flatness-Aware Learning
53. NeurIPS 2026Transductive Generalization for GNNs via Optimal Transport
54. NeurIPS 2026Transformers Can Learn Multiclass Classification In-Context: Isotropy Governs Generalization
55. NeurIPS 2026Understanding Generalization Requires Universal Induction
56. NeurIPS 2026Unlocking Compositional Generalization in Continual Few-Shot Learning
57. NeurIPS 2026Verifiable Environments Are LEGO Bricks: Recursive Composition for Reasoning Generalization
58. NeurIPS 2026WavCIL: Wavelet Coefficient-Domain Invariant Learning for Dynamic Graph OOD Generalization
59. NeurIPS 2025A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective
60. NeurIPS 2025Adversarial generalization of unfolding (model-based) networks
61. NeurIPS 2025Aggregation Hides Out-of-Distribution Generalization Failures from Spurious Correlations
62. NeurIPS 2025An Effective Levelling Paradigm for Unlabeled Scenarios
63. NeurIPS 2025Automatic Visual Instrumental Variable Learning for Confounding-Resistant Domain Generalization
64. NeurIPS 2025Balancing Positive and Negative Classification Error Rates in Positive-Unlabeled Learning
65. NeurIPS 2025Beyond Benign Overfitting in Nadaraya-Watson Interpolators
66. NeurIPS 2025Born a Transformer -- Always a Transformer? On the Effect of Pretraining on Architectural Abilities
67. NeurIPS 2025Continuous Domain Generalization
68. NeurIPS 2025Curriculum Model Merging: Harmonizing Chemical LLMs for Enhanced Cross-Task Generalization
69. NeurIPS 2025Extrapolation by Association: Length Generalization Transfer In Transformers
70. NeurIPS 2025Flatness is Necessary, Neural Collapse is Not: Rethinking Generalization via Grokking
71. NeurIPS 2025From Linear to Nonlinear: Provable Weak-to-Strong Generalization through Feature Learning
72. NeurIPS 2025Generalization Bounds for Model-based Algorithm Configuration
73. NeurIPS 2025Generalization Bounds for Rank-sparse Neural Networks
74. NeurIPS 2025Generalization Error Analysis for Selective State-Space Models Through the Lens of Attention
75. NeurIPS 2025Generalization Guarantees for Learning Score-Based Branch-and-Cut Policies in Integer Programming
76. NeurIPS 2025Generalization vs Specialization under Concept Shift
77. NeurIPS 2025Generalizing Verifiable Instruction Following
78. NeurIPS 2025Generalizing while preserving monotonicity in comparison-based preference learning models
79. NeurIPS 2025Gradient-Weight Alignment as a Train-Time Proxy for Generalization in Classification Tasks
80. NeurIPS 2025HeroFilter: Adaptive Spectral Graph Filter for Varying Heterophilic Relations
81. NeurIPS 2025How Ensembles of Distilled Policies Improve Generalisation in Reinforcement Learning
82. NeurIPS 2025How Many Domains Suffice for Domain Generalization? A Tight Characterization via the Domain Shattering Dimension
83. NeurIPS 2025How Patterns Dictate Learnability in Sequential Data
84. NeurIPS 2025Information-theoretic Generalization Analysis for VQ-VAEs: A Role of Latent Variables
85. NeurIPS 2025Is Grokking a Computational Glass Relaxation?
86. NeurIPS 2025Jamais Vu: Exposing the Generalization Gap in Supervised Semantic Correspondence
87. NeurIPS 2025Learning Robust Spectral Dynamics for Temporal Domain Generalization
88. NeurIPS 2025Learning to Generalize: An Information Perspective on Neural Processes
89. NeurIPS 2025Length Generalization via Auxiliary Tasks
90. NeurIPS 2025Less is More: Local Intrinsic Dimensions of Contextual Language Models
91. NeurIPS 2025Linear Transformers Implicitly Discover Unified Numerical Algorithms
92. NeurIPS 2025Making Classic GNNs Strong Baselines Across Varying Homophily: A Smoothness–Generalization Perspective
93. NeurIPS 2025Mamba Modulation: On the Length Generalization of Mamba Models
94. NeurIPS 2025Minimal Semantic Sufficiency Meets Unsupervised Domain Generalization
95. NeurIPS 2025Multiclass Loss Geometry Matters for Generalization of Gradient Descent in Separable Classification
96. NeurIPS 2025On Inductive Biases That Enable Generalization in Diffusion Transformers
97. NeurIPS 2025On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity
98. NeurIPS 2025On the Mechanisms of Weak-to-Strong Generalization: A Theoretical Perspective
99. NeurIPS 2025On the VC dimension of deep group convolutional neural networks
100. NeurIPS 2025On Transferring Transferability: Towards a Theory for Size Generalization

常见问题

2023–2026 年四大会议一共有多少篇通用机器学习 · generalization论文？

共 592 篇，按年份分别是 2023 年 136 篇、2024 年 135 篇、2025 年 169 篇、2026 年 152 篇。

NeurIPS 2026 有多少篇通用机器学习 · generalization论文？

NeurIPS 2026 有 58 篇通用机器学习 · generalization论文，上一届是 61 篇。

ICLR 2026 有多少篇通用机器学习 · generalization论文？

ICLR 2026 有 38 篇通用机器学习 · generalization论文，上一届是 33 篇。

ICML 2026 有多少篇通用机器学习 · generalization论文？

ICML 2026 有 47 篇通用机器学习 · generalization论文，上一届是 61 篇。

CVPR 2026 有多少篇通用机器学习 · generalization论文？

CVPR 2026 有 9 篇通用机器学习 · generalization论文，上一届是 14 篇。

哪个会议的通用机器学习 · generalization论文最多？

NeurIPS，2023–2026 年共 207 篇。