AI/CORE A*

COLT 2026

Annual Conference on Learning Theory

June 29 - July 3, 2026San Diego, CaliforniaOfficial conference site Site reachable

COLT 2026 is the 39th Annual Conference on Learning Theory, focusing on theoretical aspects of machine learning at the intersection of computer science, statistics, and applied mathematics. It welcomes submissions that advance the understanding of learnability, optimization, and empirical phenomena in learning, with an inclusive view of learning theory that spans classical and emerging domains.

Official CFP Back to deadlines Verified September 9, 2026

Key deadlines

Verified September 9, 2026

Full paper

February 4, 2026

AoE

Conference timeline

Submission and decisions

Full paperKey deadline

February 4, 2026 · AoE

Paper fit

Contribution paths

A strong submission should clearly identify its contribution and evaluate it appropriately.

Full Paper

Submissions must be a single PDF with a main body limited to 12 PMLR-formatted pages (excluding references). All proofs and details must be included, possibly in appendices, but novelty and significance are judged primarily on the main paper.

1-Page Extended Abstract

Authors of accepted papers may opt out of the full proceedings in favor of a 1-page extended abstract pointing to an open access archival version of the full paper reviewed for COLT.

Research areas in scope

Core Learning Theory Topics

Design and analysis of learning algorithmsStatistical and computational complexity of learningOptimization methods for learning, including online and stochastic optimizationTheory of artificial neural networks, including deep learningTheoretical explanation of empirical phenomena in learningSupervised learningUnsupervised, semi-supervised learning, domain adaptationLearning geometric and topological structures in data, manifold learningActive and interactive learningReinforcement learningOnline learning and decision-makingInteractions of learning theory with other mathematical fieldsHigh-dimensional and non-parametric statisticsKernel methodsCausalitySamplingTheoretical analysis of probabilistic graphical modelsBayesian methods in learningGame theory and learningLearning with system constraints (e.g., privacy, fairness, memory, communication)Learning from complex data (e.g., networks, time series)Learning in neuroscience, social science, economics and other subjectsQuantum learning theory

Policies worth checking twice

Submissions must be anonymized for double-blind review: no author names or identifying information in the submission.
The main paper is limited to 12 PMLR-formatted pages; references and appendices have no page limit.
All proofs and derivations required to substantiate results must be included in the submission.
Dual submissions to other peer-reviewed conferences with proceedings are prohibited.
Dual submissions to journals are prohibited; COLT submissions may not be submitted to a journal until after decisions are released or withdrawn.
At least one author of each accepted paper must attend the conference in person to present the work.
Authors may use large language models (LLMs) for writing, mathematics, and research, but must disclose unusual contributions and remain responsible for correctness.
Reviewers must not share submissions with any LLM.

Official sources

Compiled from the official call for papers. The organizers’ pages remain authoritative.

Official conference site Call for Papers Conference Program Workshops, Tutorials, Community Program committee Accepted Papers

Last verified September 9, 2026