ELLIS Summer School on Autonomous Driving 2026 (SSAD)
📅 Tuesday, 15 September 2026 → Friday, 18 September 2026 in 64 days
Four-day Barcelona school on end-to-end driving, world models, BEV perception and VLA models, with speakers from NVIDIA, Wayve, Valeo.ai and TU Delft.
A four-day school at the intersection of machine learning, computer vision and robotics, run by the Computer Vision Center with Valeo.ai and the ELLIS Unit Barcelona. The topic list is a fair map of where autonomous driving research actually is: end-to-end driving, 3D and BEV perception, foundation models, world models, reinforcement learning, sim-to-real, SLAM and vision-language-action models.
Nine speakers are confirmed, drawn unusually evenly from academia and the companies doing the work: José M. Álvarez (NVIDIA) on end-to-end driving and VLA models; Vassia Simaiaki (Wayve) on world models; Alexandre Boulch (Valeo.ai/Inria) on multi-sensor pre-training; German Ros (NVIDIA/CARLA) on the CARLA simulator; Maximilian Igl (NVIDIA) on RL for decision-making; Holger Caesar (TU Delft) on dataset construction; Fatma Güney (Koç) on reliable perception; Kashyap Chitta on benchmarking; and Antonio M. López (CVC/UAB).
The programme runs 09:00–17:00 across four days at UAB Casa Convalescència, with two student poster sessions, two panels, a CVC campus visit for autonomous-vehicle demos, and a planned hackathon on a 6,000km Catalunya driving dataset. Fees are €400 for PhD students and €600 for postdocs and industry. Applications have closed (extended deadline 12 June 2026); the final schedule is published around six weeks before the school.
