Main Conference Papers
Original, unpublished research papers on probabilistic graphical models, limited to 12 pages excluding references, to be presented as oral or poster presentations and published in PMLR.
AI/CORE Unranked
PGM2026 2026
The 13th International Conference on Probabilistic Graphical Models
PGM2026 is the 13th International Conference on Probabilistic Graphical Models, a biennial event bringing together researchers in artificial intelligence, machine learning, and statistics to advance probabilistic reasoning, decision making, and learning using graphical models. The conference will be held in Valencia, Spain, in September 2026, featuring a main conference and a pre-conference workshop focused on applications in Earth System Science.
Key deadlines
Verified September 9, 2026Abstract registration
May 23, 2026
AoE
Full paper
June 2, 2026
AoE
Conference timeline
Submission and decisions
Abstract registrationKey deadline
May 23, 2026 · AoE
Full paperKey deadline
June 2, 2026 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Workshop Abstracts
Non-archival abstracts (up to 500 words or 1 page) for the pre-conference workshop on Probabilistic Graphical Models for Earth System Science, to be presented orally or as posters without formal proceedings.
Research areas in scope
01
Main Conference Topics
Probabilistic graphical models (Bayesian networks, chain graphs, influence diagrams, sum–product networks, probabilistic relational models)Exact and approximate inference, Monte Carlo methods, variational approachesStructure learning, parameter learning, causal discoveryCausal modeling and causal inferenceInterpretable, explainable, and trustworthy AI based on graphical modelsDecision theory, reasoning under uncertainty, sequential decision processesConnections between graphical models and deep learning, optimization, statistical learningScalable algorithms and software systems for graphical modelsApplications in science, engineering, healthcare, economics, climate, and environmental sciences
02
Workshop Topics (PGM-ESS 2026)
Efficient exact and approximate inference algorithms for environmental modelsLearning graphical representations under data sparsity and missing data in spatiotemporal datasetsCausal discovery from high-dimensional, spatiotemporal, or non-stationary climate dataCausal inference and attribution of extreme events (e.g., heatwaves, droughts, floods)Physics-informed PGMs incorporating physical laws or numerical model outputsSpatiotemporal graphical models (dynamic Bayesian networks, state-space models)Explainable and trustworthy AI in climate scienceBayesian networks for ecological management, environmental policy, and risk assessmentAsymmetric graphical models (staged trees, context-specific independence)Uncertainty quantification in climate projections
Policies worth checking twice
- Submissions must report original, unpublished research not under review elsewhere.
- All submissions undergo a rigorous peer-review process.
- Papers must be submitted in LaTeX using the official PGM 2026 Author Kit template.
- Papers are limited to 12 pages excluding references.
- Accepted papers will be published in the Proceedings of Machine Learning Research (PMLR).
- At least one author of each accepted paper must register for the conference.
- Generative AI tools (e.g., LLMs) may be used to assist in writing or research, but authors bear full responsibility for content and must disclose notable AI usage.
- LLMs are not eligible for authorship.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026