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

540 papers, page 2 of 11

51. NeurIPS 2026Inference Time Nash Alignment
52. NeurIPS 2026Inferential Theory of Learning as a Framework for Test-Time Computation in Foundation Models
53. NeurIPS 2026InfoFlow: A Framework for Multi-Layer Transformer Analysis
54. NeurIPS 2026Information Loss and Disparate Effects in Network Embeddings
55. NeurIPS 2026Learning Acceptable Lotteries via Queries: Minimizing Aggregated Violation Distances
56. NeurIPS 2026Learning Chance-Constrained MDPs with Bellman Distributional Certificates
57. NeurIPS 2026Learning Lotteries with Minimal Violations from Membership Queries
58. NeurIPS 2026Learning to Audit ML Models with Theory of Mind
59. NeurIPS 2026Less Structure is More: Minimal Representations for Supervised Learning
60. NeurIPS 2026Linear Contextual Bandits with Quasi-Optimism
61. NeurIPS 2026Linear Ensemble Sampling with Fewer Ensembles
62. NeurIPS 2026Looped Transformers with Layer Normalization Provably Learn the Power Method
63. NeurIPS 2026Memory Determines Learning Direction: A Theory of Gradient-Based Optimization in State Space Models
64. NeurIPS 2026Minimax Optimal Early-Stopped Gradient Descent for Gaussian Mixture Classification
65. NeurIPS 2026MinMax Recurrent Neural Cascades
66. NeurIPS 2026Model Collapse is a Singular Complexity Trajectory
67. NeurIPS 2026Multi-Armed Bandits With Best-Action Queries
68. NeurIPS 2026Multiple Descent of Generalization Curve for Optimally Regularized Ridge Regression
69. NeurIPS 2026Near-Optimal Best-of-Both-Worlds Algorithms for Decoupled Exploration and Exploitation in Multi-armed Bandits
70. NeurIPS 2026Network Intervention by Polling Strategic Agents
71. NeurIPS 2026Nonparametric In-Context Learning under Growing Geometric Complexity: Minimax Optimality and Local Geometry-Adaptivity of Transformers
72. NeurIPS 2026On Concentration Inequalities for Sampling without Replacement
73. NeurIPS 2026On the Complexity of Preference-Based Bandits
74. NeurIPS 2026One-Layer Transformers Provably Learn In-Context K-Nearest Neighbor Prediction with Chain-of-Thought
75. NeurIPS 2026Online Directional Regression for Streaming Sufficient Dimension Reduction
76. NeurIPS 2026Online Resource Allocation With General Constraints
77. NeurIPS 2026Online Set Learning from Precision and Recall Feedback
78. NeurIPS 2026Optimal Dimension-Free Sampling for Regularized Classification
79. NeurIPS 2026PAC-Bayesian Bounds for Learning Partially Observed Stochastic Linear Time-Invariant State-Space Systems with Inputs and Sub-Gaussian Noise
80. NeurIPS 2026Parameter symmetries determine representational geometry in overparameterized nonlinear networks
81. NeurIPS 2026Phase Transitions in Attention: A Bayesian Theory of Copy Head Emergence
82. NeurIPS 2026Policy Optimization in Tabular MDPs: Data-dependent Regret under Unknown Transitions
83. NeurIPS 2026Post-Processing Guarantees for Classification under Linear-Fractional Performance Metrics
84. NeurIPS 2026Prediction Under Imperfect Compression: A Theory of Approximate MDL
85. NeurIPS 2026Propagation of Chaos in Contextual Flow Maps
86. NeurIPS 2026Random Matrix Theory of Early-Stopped Gradient Flow: A Transient BBP Scenario
87. NeurIPS 2026Regret Minimization in Single-Dimensional Contract-Design with Binary Actions
88. NeurIPS 2026Replicable Constrained Bandits
89. NeurIPS 2026Resolving AdaBoost Cycling with LLMs: A Computer-Assisted Counterexample
90. NeurIPS 2026Robust Statistical Estimators with Bounded Empirical Sensitivity
91. NeurIPS 2026Safe Linear Bandits with Unknown Safety Gaps
92. NeurIPS 2026SAMPPO: Structure-Aware Mirror Proximal Policy Optimization
93. NeurIPS 2026Second-Order Complexity Theory for Neural Networks
94. NeurIPS 2026Selling Information While Being an Interested Party
95. NeurIPS 2026Sharp Non-Asymptotic Analysis of the Penalized Challenger in $\beta$-EB-TCI for Bernoulli Bandits
96. NeurIPS 2026Testable and Actionable Calibration for Full Swap Regret
97. NeurIPS 2026The Reflexivity Threshold: A Phase Transition for Multi-Agent Performative Prediction
98. NeurIPS 2026The Symmetries of Three-Layer ReLU Networks
99. NeurIPS 2026Theory of Optimal Learning Rate Schedules and Scaling Laws for a Random Feature Model
100. NeurIPS 2026Theory on Attention Dynamics for Out-of-Distribution In-Context Learning

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.