Research papers
Focus on research work, datasets, benchmarks, competitions, systems, deployment experiences, or position papers that address real-world deployment challenges and move beyond idealised benchmark conditions.
Systems/CORE Unranked
BMVC26 MMAIRS 2026
BMVC 2026 Workshop on Multimodal AI for real-world Systems
The BMVC Workshop on Deployable AI for Real-World Systems focuses on multimodal machine learning and computer vision for AI systems that operate beyond controlled environments, addressing real-world challenges such as noisy data, missing modalities, and resource constraints. It welcomes research on algorithms, datasets, benchmarks, and deployment case studies across domains like robotics, healthcare, autonomous vehicles, and edge AI, with a special emphasis on moving beyond idealized benchmarks.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Research areas in scope
01
Topics of Interest
Multimodal learning for real-world and deployed systemsComputer vision and multimodal perception under deployment constraintsSensor fusion for autonomous systems, robotics, transport and smart infrastructureMultimodal AI beyond vision and languageLearning from incomplete, noisy, weakly labelled or missing modalitiesCross-modal alignment, synchronisation and calibrationRobustness to distribution shift, sensor failure, weather, occlusion and domain changeMultimodal learning for edge, embedded and resource-constrained devicesNeural Architecture Search, AutoML and hardware-aware model design for multimodal systemsEfficient multimodal model selection, compression, distillation and deploymentMultimodal datasets, benchmarks and evaluation protocols for real-world scenariosMeasuring the contribution, redundancy and complementarity of different modalitiesSafety, reliability, uncertainty estimation and failure detection in multimodal AIInterpretability and explainability for multimodal decision-makingPrivacy, ethics, fairness and governance in deployed multimodal systemsHuman-in-the-loop multimodal AI and user-facing deployment workflowsApplications in autonomous vehicles, transport, robotics, healthcare, environmental monitoring, industrial inspection, biodiversity, agriculture, sustainability and public servicesCase studies, lessons learned and negative results from real-world multimodal AI deployments
Policies worth checking twice
- Papers are limited to eight pages, excluding references and appendices.
- Papers will be published non-archivally.
- Submissions must be made via OpenReview.
- Submissions should clearly explain the real-world setting or deployment challenge being addressed.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026