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

1,350 papers, page 2 of 27

51. ICLR 2026Efficient Turing Machine Simulation with Transformers
52. ICLR 2026Emergence of Superposition: Unveiling the Training Dynamics of Chain of Continuous Thought
53. ICLR 2026Escaping Model Collapse via Synthetic Data Verification: Near-term Improvements and Long-term Convergence
54. ICLR 2026FACT: a first-principles alternative to the Neural Feature Ansatz for how networks learn representations
55. ICLR 2026Fast Catch-Up, Late Switching: Optimal Batch Size Scheduling via Functional Scaling Laws
56. ICLR 2026Fast Escape, Slow Convergence: Learning Dynamics of Phase Retrieval under Power-Law Data
57. ICLR 2026Faster Parameter-Free Regret Matching Algorithms
58. ICLR 2026Feature compression is the root cause of adversarial fragility in neural networks
59. ICLR 2026Feedback-driven recurrent quantum neural network universality
60. ICLR 2026Finite-Time Analysis of Actor-Critic Methods with Deep Neural Network Approximation
61. ICLR 2026Frozen Policy Iteration: Computationally Efficient RL under Linear $Q^{\pi}$ Realizability for Deterministic Dynamics
62. ICLR 2026GenCtrl -- A Formal Controllability Toolkit for Generative Models
63. ICLR 2026Gradient Descent Dynamics of Rank-One Matrix Denoising
64. ICLR 2026Group Representational Position Encoding
65. ICLR 2026High Probability Bounds for Non-Convex Stochastic Optimization with Momentum
66. ICLR 2026High-Dimensional Analysis of Single-Layer Attention for Sparse-Token Classification
67. ICLR 2026High-dimensional limit theorems for SGD: Momentum and Adaptive Step-sizes
68. ICLR 2026How reinforcement learning after next-token prediction facilitates learning
69. ICLR 2026Implicit Models: Expressive Power Scales with Test-Time Compute
70. ICLR 2026Implicit Sensing for Fourier Sparse Boolean Functions
71. ICLR 2026Improved high-dimensional estimation with Langevin dynamics and stochastic weight averaging
72. ICLR 2026Incentives in Federated Learning with Heterogeneous Agents
73. ICLR 2026Infinite Horizon Markov Economies
74. ICLR 2026InfoNCE Induces Gaussian Distribution
75. ICLR 2026Information Estimation with Discrete Diffusion
76. ICLR 2026InfoScan: Information-Efficient Visual Scanning via Resource-Adaptive Walks
77. ICLR 2026InfoTok: Adaptive Discrete Video Tokenizer via Information-Theoretic Compression
78. ICLR 2026Inheriting Generalizable Knowledge from LLMs to Diverse Vertical Tasks
79. ICLR 2026Instance-Dependent Fixed-Budget Pure Exploration in Reinforcement Learning
80. ICLR 2026Interactive Learning of Single-Index Models via Stochastic Gradient Descent
81. ICLR 2026Intrinsic training dynamics of deep neural networks
82. ICLR 2026Language Identification in the Limit with Computational Trace
83. ICLR 2026Language Models Use Lookbacks to Track Beliefs
84. ICLR 2026Larger Datasets Can Be Repeated More: A Theoretical Analysis of Multi-Epoch Scaling in Linear Regression
85. ICLR 2026Learning a Game by Paying the Agents
86. ICLR 2026Learning in Prophet Inequalities with Noisy Observations
87. ICLR 2026Learning is Forgetting; LLM Training As Lossy Compression
88. ICLR 2026Learning on a Razor’s Edge: Identifiability and Singularity of Polynomial Neural Networks
89. ICLR 2026Learning Shrinks the Hard Tail: Training‑Dependent Inference Scaling in a Solvable Linear Model
90. ICLR 2026Learning the Inverse Temperature of Ising Models under Hard Constraints using One Sample
91. ICLR 2026Learning to Answer from Correct Demonstrations
92. ICLR 2026Learning to Play Multi-Follower Bayesian Stackelberg Games
93. ICLR 2026Learning to Recall with Transformers Beyond Orthogonal Embeddings
94. ICLR 2026Learning under Quantization for High-Dimensional Linear Regression
95. ICLR 2026Learning-Augmented Moment Estimation on Time-Decay Models
96. ICLR 2026Mean Estimation from Coarse Data: Characterizations and Efficient Algorithms
97. ICLR 2026Memorizing Long-tail Data Can Help Generalization Through Composition
98. ICLR 2026Memory-Statistics Tradeoff in Continual Learning with Structural Regularization
99. ICLR 2026Metric $k$-clustering using only Weak Comparison Oracles
100. ICLR 2026Minimax Sample Complexity of Graph Neural Networks: Lower Bounds and Structural Effects

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.