Archival papers
8-14 pages long excluding references; accepted papers will be included in the ECCV workshop proceedings.
Computer Vision/CORE Unranked
WMEAI 2026
Workshop on How to Build Effective World Models for Embodied AI
The WMEAI 2026 workshop, held in conjunction with ECCV 2026 in Malmö, Sweden, brings together researchers from computer vision, robotics, language, and generative modeling to explore the development and integration of world models in embodied AI. The workshop focuses on key challenges such as physical representation learning, simulation-to-real transfer, evaluation protocols, and scaling world models for real-world deployment.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Non-archival papers
Limited to 7 pages excluding references.
Research areas in scope
01
Topics of interest
World models integration into embodied AI systemsData generation, training, and optimizationSimulation, imagination, and generative latent modelsPhysical and spatial representation learningEvaluation protocols and benchmarkingLearned world models for planning and controlScaling world models across different domainsSim-to-real transfer for real-world deployment
Policies worth checking twice
- Submissions must follow the official ECCV'26 guidelines
- Submissions are evaluated based on novelty, relevance to workshop topics, and significance
- Accepted archival papers will be included in the ECCV workshop proceedings
- Submissions are handled via OpenReview
- Camera-ready deadline is August 15, 2026
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026