Regular Papers
Full-length research papers presenting original contributions in algorithmic learning theory.
AI/CORE B
ALT 2026
International Conference on Algorithmic Learning Theory
ALT 2026 is the 37th International Conference on Algorithmic Learning Theory, held in Toronto, Canada, alongside ShaiFest—a special event honoring Shai Ben-David. The conference showcases cutting-edge research in algorithmic learning theory, covering theoretical and practical advances in machine learning, online learning, statistical learning, and related areas.
Key deadlines
Verified September 9, 2026Full paper
October 2, 2025
AoE
Conference timeline
Submission and decisions
Full paperKey deadline
October 2, 2025 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Distinguished Papers
Selected papers recognized for exceptional quality and impact, presented as distinguished contributions.
Research areas in scope
01
Research Topics
Algorithmic Learning TheoryOnline LearningStatistical LearningDifferential PrivacyConformal PredictionReinforcement LearningBandit AlgorithmsOptimization MethodsDistribution LearningSample ComplexityGeneralization BoundsNeural Networks and Deep LearningLearning with Adversarial CorruptionsCovariate ShiftMulti-distribution LearningLearning from Synthetic DataData VisualizationGame Theory and LearningLearning Theory for Time SeriesQuery-Based LearningFairness and Robustness in LearningNonparametric LearningConvex and Non-convex OptimizationHigh-dimensional LearningContinual LearningLearning with Limited FeedbackLearning with Monotone AdversariesLearning with Graphs and NetworksInformation-Theoretic LearningLearning with ConstraintsParameter-Efficient LearningLearning from Distributed DataLearning with Noisy ObservationsLearning with Structured OutputsLearning with Human-in-the-LoopLearning with Causal StructureLearning with Non-stationary EnvironmentsLearning with Sparse RepresentationsLearning with Kernel MethodsLearning with Probabilistic ModelsLearning with Uncertainty QuantificationLearning with Active SamplingLearning with Transfer and Meta-LearningLearning with Privacy GuaranteesLearning with Computational ConstraintsLearning with Non-i.i.d. DataLearning with Partial ObservabilityLearning with Non-convex LossesLearning with Non-Euclidean GeometryLearning with High Probability BoundsLearning with Adaptive AlgorithmsLearning with Implicit RegularizationLearning with Model CompressionLearning with Dynamic EnvironmentsLearning with Temporal DependenciesLearning with Graphical ModelsLearning with Causal InferenceLearning with Multi-objective OptimizationLearning with Human PreferencesLearning with Fairness ConstraintsLearning with Robustness GuaranteesLearning with ExplainabilityLearning with Causal DiscoveryLearning with Domain AdaptationLearning with Out-of-Distribution GeneralizationLearning with Distributional RobustnessLearning with Adversarial ExamplesLearning with Data PoisoningLearning with Model InterpretabilityLearning with Uncertainty PropagationLearning with Bayesian MethodsLearning with Ensemble MethodsLearning with BoostingLearning with Sampling TechniquesLearning with Variance ReductionLearning with Stochastic OptimizationLearning with Mirror DescentLearning with Projected Gradient MethodsLearning with Non-smooth ObjectivesLearning with Subdifferential AnalysisLearning with Convergence AnalysisLearning with Regret BoundsLearning with Minimax OptimalityLearning with Lexicographic OptimizationLearning with Entropic RegularizationLearning with Optimal TransportLearning with Tyler’s M-EstimatorLearning with Covariance EstimationLearning with Discrete DistributionsLearning with Regular ExpressionsLearning with Decision ListsLearning with HypergraphsLearning with Graph InferenceLearning with Kernel Two-Sample TestsLearning with Large Average Subtensor ProblemsLearning with Planted PartitioningLearning with Proximal Point MethodsLearning with Adaptive Step SizesLearning with Adam OptimizationLearning with Purely Private EstimationLearning with Discrepancy TheoryLearning with Information GainLearning with Gaussian ProcessesLearning with Effective Resistance QueriesLearning with Martingale MethodsLearning with LIL RegretLearning with Graphical ModelsLearning with Algorithmic Complexity MeasuresLearning with Representation LearningLearning with Lifelong LearningLearning with Multi-task LearningLearning with RegularizationLearning with High-dimensional RegressionLearning with Non-uniform GenerationLearning with Multiple ExpertsLearning with Scale-sensitive DimensionsLearning with Distribution ShiftLearning with Generalization TheoryLearning with Uniform ConvergenceLearning with Glivenko-Cantelli ClassesLearning with PAC-Bayesian BoundsLearning with Discriminative FeedbackLearning with Teacher ClassesLearning with Replicable BoostingLearning with Pareto-Optimal GenerationLearning with Data-Dependent ParadigmsLearning with No Scale Sensitive DimensionLearning with Algorithmic MeasuresLearning with Optimal RegularizationLearning with Adam Parameter TuningLearning with Covariance EstimationLearning with Distribution LearningLearning with Distribution Shift TheoryLearning with Robustness Theory
Policies worth checking twice
- Submissions must be original and not simultaneously submitted to another conference or journal.
- Papers must be anonymized for double-blind review.
- Page limit is not explicitly stated but implied by accepted paper lengths (typically 15–25 pages).
- Author lists cannot be changed after the submission deadline.
- AI-generated content must be disclosed and used in accordance with ethical guidelines.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026