Full-length papers
Archival papers limited to 14 pages (excluding references), to be included in the ECCV workshop proceedings.
Computer Vision/CORE Unranked
EmergingAD 2026
ECCV 2026 Workshop on Emerging Behaviors for Achieving Robust Autonomy [archival]
EmergingAD 2026 is an ECCV 2026 workshop focused on emerging behaviors in embodied AI for achieving robust autonomy in autonomous driving. It brings together researchers to explore advancements in self-play reinforcement learning, imitation learning, generative world models, and Vision-Language-Action (VLA) systems, with an emphasis on closing the simulation-to-reality gap and enabling safe, human-like decision-making.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Short papers
Non-archival papers limited to 7 pages (excluding references).
Research areas in scope
01
Topics of Interest
Self-Play & Multi-Agent RL for DrivingImitation Learning, Offline RL & Online Fine-TuningEmergent Skills & Human-Like BehaviorsVision-Language-Action Models for DrivingGenerative World Models & Controllable SimulationRL for Behavior & RL for ReasoningGrounding Self-Play Policies with Real-World DataFailure Modes, Robustness & Deployment LimitationsVLA Integration with Fast SimulatorsLarge Models in Real-Time Simulation
Policies worth checking twice
- Submissions must follow the official ECCV'26 guidelines.
- Submissions are evaluated on novelty, relevance to workshop topics, technical clarity, and overall significance.
- Accepted full-length papers will be included in the ECCV workshop proceedings.
- Accepted papers will be presented as posters; two selected papers will receive 15-minute oral presentations.
- Submissions must be made via OpenReview (separate portals for archival and non-archival papers).
- Author registration deadline is 10.08.2026.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026