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

482 papers, page 8 of 10

351. ICML 2026Theoretical Perspectives on Data Quality and Synergistic Effects in Pre- and Post-Training Reasoning Models
352. ICML 2026Tight Margin-Based Generalization Bounds for Voting Classifiers over Finite Hypothesis Sets
353. ICML 2026Tight Stability Bounds for Robust Distributed Learning: Byzantine Failures Hurt Generalization More than Data Poisoning
354. ICML 2026To Grok Grokking: Provable Grokking in Ridge Regression
355. ICML 2026Token Sample Complexity of Attention
356. ICML 2026Towards Atoms of Large Language Models
357. ICML 2026Towards Spectroscopy: Susceptibility Clusters in Language Models
358. ICML 2026Towards Theoretical Understanding of Transformer Test-Time Computing: Investigation on In-Context Linear Regression
359. ICML 2026Towards Understanding Generalization of Federated Adversarial Learning: Perspective of Algorithmic Stability
360. ICML 2026Trajectory Seriation via Spectral Tangent Alignment and Global Embedding
361. ICML 2026Transformers Efficiently Perform In-Context Logistic Regression via Normalized Gradient Descent
362. ICML 2026Understanding Generalization from Embedding Dimension and Distributional Convergence
363. ICML 2026Understanding Private Learning From Feature Perspective
364. ICML 2026Unifying Low Dimensional Spectra in Deep Learning
365. ICML 2026Universal Learning of Nonlinear Dynamics
366. ICML 2026Universal Multiclass Transductive Online Learning
367. ICML 2026What Does Preference Learning Recover from Pairwise Comparison Data?
368. ICML 2026Why Agentic Theorem Prover Works: A Statistical Provability Theory of Mathematical Reasoning Models
369. ICML 2026Why DDIM Hallucinates More Than DDPM: A Theoretical Analysis of Reverse Dynamics
370. ICML 2026Why Self-Distillation Helps and Hurts: Denoising vs. Signal Forgetting
371. NeurIPS 2026A dynamical systems theory of reward-modulated learning in linear recurrent networks
372. NeurIPS 2026A Mechanistic Investigation of Theory of Mind in a Large Language Model
373. NeurIPS 2026A Near-optimal SQ Lower Bound for Smoothed Agnostic Learning of Boolean Halfspaces
374. NeurIPS 2026A Statistical Theory of Gated Attention through the Lens of Hierarchical Mixture of Experts
375. NeurIPS 2026A Theory of Adversary-Directed Online Learning
376. NeurIPS 2026A Theory of Spatial Continuous Attractors in Hopfield Energy Landscapes
377. NeurIPS 2026A Theory of Training Profit-Optimal LLMs
378. NeurIPS 2026A Theory on Flow Matching with Neural Networks
379. NeurIPS 2026A Unified Spectral Theory of Multimodal Losses
380. NeurIPS 2026Adaptive Calibration in Non-Stationary Environments
381. NeurIPS 2026An exact information theory of generalization phase transitions in Bayesian diffusion models
382. NeurIPS 2026Approximate Envy-Free Allocations up to any k Goods
383. NeurIPS 2026Autoregressive Learning in Joint KL: Sharp Oracle Bounds and Lower Bounds
384. NeurIPS 2026Avoiding Obfuscation with Prover-Estimator Debate
385. NeurIPS 2026Best Arm Identification for Bandits with Shifting Means
386. NeurIPS 2026Bilinear Matching Bandits
387. NeurIPS 2026Block Optimism for Nonstationary Bandits with Latent Linear Dynamics
388. NeurIPS 2026Borda-Based Fair Multi-User Dueling Bandit in Tabular and Generalized Linear Settings
389. NeurIPS 2026Canonical Predictive Quotients: A Theory of Prediction under Hidden Predictive State Uncertainty with ICL Implications
390. NeurIPS 2026Chain of Dual Structures in Transformer Attention
391. NeurIPS 2026Characterizing Learning in Deep Neural Networks using a Tractable Algorithmic Complexity Estimator
392. NeurIPS 2026Characterizing Trainability of Instantaneous Quantum Polynomial Circuit Born Machine
393. NeurIPS 2026Constrained MDPs with Trajectory Constraints
394. NeurIPS 2026Curvature-Dependent Lower Bounds for Riemannian Online Convex Optimization
395. NeurIPS 2026Data-Free Metrics Are Not Invariant Under Functionality-Preserving Reparametrisations
396. NeurIPS 2026Decentralized Coupled Representation Learning
397. NeurIPS 2026Diagonalizing the Softmax: Hadamard Initialization for Tractable Cross-Entropy Dynamics
398. NeurIPS 2026Diffusion Models Observe Only Gradients: A Geometric Perspective on Score Matching Errors
399. NeurIPS 2026Direction-Aware Offline-to-Online Learning in Linear Contextual Bandits
400. NeurIPS 2026Dynamic Optimistic Constrained OCO with Memory via Delay Equivalence

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.