Theory · learning theory papers, 2023–2026

Key findings

• Theory · learning theory went from 86 papers in 2023 to 482 in 2026, roughly 5.6x in three years.
• NeurIPS publishes the most Theory · learning theory work: 540 papers from 2023 to 2026, 40% of the four-venue total.
• The fastest-growing overlap is with Social aspects · accountability transparency and interpretability: papers tagged with both rose from 0 in 2023 to 33 in 2026.
• Among the 5 topics under Theory and Statistics, it ranks #2 by 2026 paper count; the fastest grower in the group, Theory · optimization, is up 4.7x versus 5.6x here.

Representative papers

Spread across years and venues so every edition shows up

• NeurIPS 2026A dynamical systems theory of reward-modulated learning in linear recurrent networks
• ICML 2026A Fine-Grained Understanding of Uniform Convergence for Halfspaces
• ICLR 2026A Law of Data Reconstruction for Random Features (And Beyond)
• NeurIPS 2025A Geometric Analysis of PCA
• ICML 2025A Generalization Theory for Zero-Shot Prediction
• ICLR 2025A Solvable Attention for Neural Scaling Laws
• CVPR 2025Dynamic Stereotype Theory Induced Micro-expression Recognition with Oriented Deformation
• NeurIPS 20244+3 Phases of Compute-Optimal Neural Scaling Laws
• ICML 2024A Statistical Theory of Regularization-Based Continual Learning
• ICLR 2024A Sublinear Adversarial Training Algorithm
• CVPR 2024Accurate Training Data for Occupancy Map Prediction in Automated Driving Using Evidence Theory
• NeurIPS 2023$\varepsilon$-fractional core stability in Hedonic Games.
• ICML 2023Robust Collaborative Learning with Linear Gradient Overhead
• ICLR 2023Bispectral Neural Networks
• NeurIPS 2026A Mechanistic Investigation of Theory of Mind in a Large Language Model
• ICML 2026A Random Matrix Theory of Masked Self-Supervised Learning
• ICLR 2026A New Approach to Controlling Linear Dynamical Systems
• NeurIPS 2025Approximation theory for 1-Lipschitz ResNets
• ICML 2025A Two-Stage Learning-to-Defer Approach for Multi-Task Learning
• ICLR 2025A Theoretical Analysis of Self-Supervised Learning for Vision Transformers

All Theory · learning theory papers

106 papers, page 2 of 3

51. ICLR 2024Learning Thresholds with Latent Values and Censored Feedback
52. ICLR 2024Learning to Reject Meets Long-tail Learning
53. ICLR 2024Learning to Relax: Setting Solver Parameters Across a Sequence of Linear System Instances
54. ICLR 2024lpNTK: Better Generalisation with Less Data via Sample Interaction During Learning
55. ICLR 2024Maximum Likelihood Estimation is All You Need for Well-Specified Covariate Shift
56. ICLR 2024Memorization Capacity of Multi-Head Attention in Transformers
57. ICLR 2024Minimax optimality of convolutional neural networks for infinite dimensional input-output problems and separation from kernel methods
58. ICLR 2024Minimum width for universal approximation using ReLU networks on compact domain
59. ICLR 2024More is Better: when Infinite Overparameterization is Optimal and Overfitting is Obligatory
60. ICLR 2024Multimodal Learning Without Labeled Multimodal Data: Guarantees and Applications
61. ICLR 2024Nearly $d$-Linear Convergence Bounds for Diffusion Models via Stochastic Localization
62. ICLR 2024Noisy Interpolation Learning with Shallow Univariate ReLU Networks
63. ICLR 2024On Representation Complexity of Model-based and Model-free Reinforcement Learning
64. ICLR 2024On the Analysis of GAN-based Image-to-Image Translation with Gaussian Noise Injection
65. ICLR 2024On the hardness of learning under symmetries
66. ICLR 2024On the Limitations of Temperature Scaling for Distributions with Overlaps
67. ICLR 2024On the Provable Advantage of Unsupervised Pretraining
68. ICLR 2024One Step of Gradient Descent is Provably the Optimal In-Context Learner with One Layer of Linear Self-Attention
69. ICLR 2024Optimal criterion for feature learning of two-layer linear neural network in high dimensional interpolation regime
70. ICLR 2024OPTIMAL ROBUST MEMORIZATION WITH RELU NEURAL NETWORKS
71. ICLR 2024Optimal Sample Complexity of Contrastive Learning
72. ICLR 2024Oracle Efficient Algorithms for Groupwise Regret
73. ICLR 2024Out-Of-Domain Unlabeled Data Improves Generalization
74. ICLR 2024Plugin estimators for selective classification with out-of-distribution detection
75. ICLR 2024Provable Memory Efficient Self-Play Algorithm for Model-free Reinforcement Learning
76. ICLR 2024Provable Offline Preference-Based Reinforcement Learning
77. ICLR 2024Quantifying the Sensitivity of Inverse Reinforcement Learning to Misspecification
78. ICLR 2024Random Sparse Lifts: Construction, Analysis and Convergence of finite sparse networks
79. ICLR 2024Rethinking Information-theoretic Generalization: Loss Entropy Induced PAC Bounds
80. ICLR 2024Risk Bounds of Accelerated SGD for Overparameterized Linear Regression
81. ICLR 2024Role of Locality and Weight Sharing in Image-Based Tasks: A Sample Complexity Separation between CNNs, LCNs, and FCNs
82. ICLR 2024Sample-Efficient Linear Representation Learning from Non-IID Non-Isotropic Data
83. ICLR 2024Sample-Efficient Multi-Agent RL: An Optimization Perspective
84. ICLR 2024Sampling Multimodal Distributions with the Vanilla Score: Benefits of Data-Based Initialization
85. ICLR 2024Scaling Laws for Associative Memories
86. ICLR 2024SGD Finds then Tunes Features in Two-Layer Neural Networks with near-Optimal Sample Complexity: A Case Study in the XOR problem
87. ICLR 2024Simplifying Transformer Blocks
88. ICLR 2024Smooth ECE: Principled Reliability Diagrams via Kernel Smoothing
89. ICLR 2024Some Fundamental Aspects about Lipschitz Continuity of Neural Networks
90. ICLR 2024Spectrally Transformed Kernel Regression
91. ICLR 2024Statistical Perspective of Top-K Sparse Softmax Gating Mixture of Experts
92. ICLR 2024Stochastic Modified Equations and Dynamics of Dropout Algorithm
93. ICLR 2024Symmetric Single Index Learning
94. ICLR 2024The Effective Horizon Explains Deep RL Performance in Stochastic Environments
95. ICLR 2024The Expressive Power of Low-Rank Adaptation
96. ICLR 2024The Human-AI Substitution game: active learning from a strategic labeler
97. ICLR 2024The optimality of kernel classifiers in Sobolev space
98. ICLR 2024Theoretical Understanding of Learning from Adversarial Perturbations
99. ICLR 2024Tight Rates in Supervised Outlier Transfer Learning
100. ICLR 2024Toward Student-oriented Teacher Network Training for Knowledge Distillation

FAQ

How many Theory · learning theory papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

1,350 in total: 86 in 2023, 393 in 2024, 389 in 2025, 482 in 2026.

How many Theory · learning theory papers are at NeurIPS 2026?

NeurIPS 2026 has 112 Theory · learning theory papers, versus 106 the year before.

How many Theory · learning theory papers are at ICLR 2026?

ICLR 2026 has 200 Theory · learning theory papers, versus 147 the year before.

How many Theory · learning theory papers are at ICML 2026?

ICML 2026 has 170 Theory · learning theory papers, versus 132 the year before.

How many Theory · learning theory papers are at CVPR 2025?

CVPR 2025 has 4 Theory · learning theory papers, versus 1 the year before.

Which conference publishes the most Theory · learning theory papers?

NeurIPS, with 540 papers from 2023 to 2026.