Written and maintained by CASRAI Editorial Board
Last updated
Target audience and objective: This guide is for machine learning researchers, mathematical statisticians, and academic authors preparing to submit papers, navigate peer review, or evaluate proceedings from the International Conference on Artificial Intelligence and Statistics (AISTATS).
What AISTATS Is
The International Conference on Artificial Intelligence and Statistics (commonly abbreviated as AISTATS) is an annual, peer-reviewed scientific conference dedicated to the mathematical, methodological, and computational intersection of statistical theory and artificial intelligence. While major computer science venues often prioritize empirical software benchmarks or engineering heuristics, AISTATS focuses on rigorous statistical foundations, theoretical guarantees, and probabilistic modelling underlying modern machine learning methodologies.
Established as an independent academic forum in 1985, AISTATS serves as a primary venue for presenting fundamental advances in statistical machine learning, Bayesian inference, causal discovery, optimization theory, and high-dimensional data analysis. In the broader taxonomy of machine learning and artificial intelligence, the conference provides an essential bridge between mathematical statistics departments and computational artificial intelligence laboratories, ensuring that newly developed learning algorithms are supported by sound inferential guarantees.
Who Runs It and What It Covers
AISTATS operates under the scientific leadership of rotating General Chairs and Program Chairs drawn from leading international universities and research laboratories. The conference operates as a non-profit academic symposium, partnering with the Society for Artificial Intelligence and Statistics to maintain institutional governance. Accepted papers are published open-access in the Proceedings of Machine Learning Research (PMLR) series, guaranteeing immediate, un-paywalled global distribution under Creative Commons licensing.
The substantive scope of AISTATS encompasses both foundational mathematical theory and applied probabilistic algorithms. Primary topical tracks include:
- Probabilistic and Bayesian Methods: Variational inference, Markov chain Monte Carlo (MCMC), Gaussian processes, and probabilistic graphical models.
- Statistical Learning Theory: Generalization bounds, non-parametric statistics, high-dimensional estimation, and minimax rates of convergence.
- Causality and Structural Equation Modeling: Causal effect estimation, observational identifiability, and counterfactual inference.
- Optimization and Computation: Stochastic convex and non-convex optimization, accelerated gradient schemes, and saddle-point dynamics.
- Modern Deep Learning Foundations: Implicit regularization, neural tangent kernels, and representation learning dynamics analyzed through statistical lenses.
According to the Computing Research and Education Association of Australasia (CORE), AISTATS is classified as a prestigious Rank A conference. Its competitive acceptance rates typically range between 27% and 31% across recent conference cycles, reflecting rigorous editorial standards comparable to leading specialty venues in computational science.
How to Submit and Review for It
Participating in AISTATS—whether as an author submitting original research or as a technical reviewer evaluating manuscripts—follows a structured protocol administered through modern academic peer review infrastructure:
- Review Submission Deadlines and Parameters: The official call for papers is issued on
https://aistats.org, typically establishing paper submission deadlines in the autumn preceding the spring conference. Authors preparing manuscripts for future cycles, such as AISTATS 2026 or AISTATS 2027, must consult official announcements for exact calendar deadlines. - Adhere to Page Limits and Formatting: Submissions must use the official AISTATS LaTeX template. Initial submissions are strictly limited to 8 pages of main text. References, reproducibility checklists, AI use disclosures, and supplementary appendices do not count toward this 8-page limitation. Accepted camera-ready papers receive an additional page (up to 9 pages of main text) to accommodate reviewer feedback.
- Maintain Strict Anonymity for Double-Blind Review: Author identities, institutional affiliations, acknowledgments, and direct funding references must be removed from the manuscript and supplementary files. Self-citations must be written in the third person.
- Submit via OpenReview: Papers are deposited through the conference’s designated OpenReview management portal. Authors specify primary and secondary subject areas to facilitate automated paper-reviewer matching based on subject affinity and bidding profiles.
- Participate in Author Rebuttal: Following initial reviewer evaluations, authors receive technical reviews and are allotted a defined window to submit a formal rebuttal addressing factual errors, clarifying proofs, or providing supplementary ablation results.
- Prepare Camera-Ready Delivery for PMLR: Authors of accepted papers finalize their manuscripts following official editorial specifications, generate reproducible source packages, and deposit camera-ready PDF files for open-access publishing in the corresponding PMLR volume.
- Index and Disseminate Code Artefacts: To ensure empirical reproducibility, authors are encouraged to link verified code repositories on community indexing platforms such as Papers with Code or deposit model weights and datasets on open hubs.
Strengths and Limits
Understanding where AISTATS sits within the scholarly publishing ecosystem helps researchers choose the most effective venue for their findings.
Key Strengths
- Mathematical and Statistical Rigor: Unlike purely empirical venues, AISTATS reviewers prioritize formal mathematical proofs, precise theorem statements, and theoretical convergence analyses.
- Archival Open-Access Publishing: Publishing in PMLR ensures that accepted research is free to read, permanent, and indexed in major scientific search engines without paywalls or author-facing article processing charges.
- High Academic Standing: A consistent CORE Rank A designation establishes substantial institutional recognition for tenure, promotion, and grant reporting across international research institutions.
- Interdisciplinary Community: Fosters genuine collaboration between traditional mathematical statisticians and computational AI researchers.
Technical and Operational Limitations
- Theoretical Barrier to Entry: Papers focused purely on large-scale engineering feats, hardware benchmarks, or practical application deployments without theoretical or statistical contributions often receive critical reviews.
- Competitive Selectivity: With acceptance rates hovering around 28%, many mathematically sound manuscripts are rejected due to limited presentation slots.
- Smaller Scale Than Broad AI Megaconferences: While prestigious, AISTATS remains more specialized and smaller in attendance than broad generalist conferences like NeurIPS or ICML.
Alternatives
Researchers evaluating publication targets in statistical learning, artificial intelligence, and uncertainty quantification can compare AISTATS against related top-tier venues:
| Conference | Primary Emphasis | Review Model | Proceedings Publisher | CORE Ranking |
|---|---|---|---|---|
| AISTATS | Statistical machine learning, Bayesian inference, probabilistic theory | Double-blind with rebuttal (OpenReview) | PMLR (Open Access) | Rank A |
| NeurIPS | Broad artificial intelligence, deep learning, neuroscience | Double-blind with rebuttal (OpenReview) | Curran Associates / NeurIPS Proceedings | Rank A* |
| ICML | General machine learning theory and computational algorithms | Double-blind with rebuttal (OpenReview) | PMLR (Open Access) | Rank A* |
| ICLR | Deep representation learning, foundation models, architectures | Double-blind open review (OpenReview) | OpenReview (Open Access) | Rank A* |
| UAI | Uncertainty in artificial intelligence, causal reasoning, decision theory | Double-blind with rebuttal | PMLR (Open Access) | Rank A |
Frequently Asked Questions
Is AISTATS a top conference?
Yes. AISTATS is internationally recognized as a premier Tier-1 venue for statistical machine learning. It holds an official Rank A rating from the CORE Conference Ranking system and maintains competitive acceptance rates between 27% and 31%.
How do I cancel my AI stats subscription?
The academic research conference AISTATS does not offer consumer subscriptions, charges no recurring user fees, and provides all proceedings completely free of charge via PMLR. If you see recurring subscription charges labeled “AI stats” or “AIstats” on your bank statement, they originate from unrelated commercial sports prediction or football analytics mobile applications (such as AIstats Inc. on Google Play or iOS). To cancel those charges, manage your subscriptions directly through the Apple App Store, Google Play Store, or the respective commercial app developer.
How does peer review work at AISTATS?
AISTATS utilizes a double-blind peer review process administered through OpenReview. Each submission is assigned to multiple expert reviewers and a Meta-Reviewer (Area Chair). After initial reviews are released, authors can submit a rebuttal to clarify technical points before final acceptance decisions are determined.
How are statistics related to AI at AISTATS?
Statistics provides the mathematical framework for understanding uncertainty, estimation error, sample complexity, and hypothesis testing in artificial intelligence. At AISTATS, papers examine how statistical principles validate machine learning models, ensuring reliable inference and generalization beyond empirical training datasets.
Sources
- AISTATS Official Conference Organization (read 2026-11-07): International Conference on Artificial Intelligence and Statistics. Available at:
https://aistats.org - Proceedings of Machine Learning Research (read 2026-11-07): PMLR AISTATS Proceedings Archive. Available at:
https://proceedings.mlr.press - CORE Portal (read 2026-11-07): Computing Research and Education Association Conference Rankings. Available at:
http://portal.core.edu.au
