The ELLIS Multimodal Learning Systems (MLS) Program held its annual meeting on July 2–3, 2026 in Vall de Núria, in the Catalan Pyrenees, bringing together 19 researchers from across the ELLIS network for two days of scientific exchange, collaboration and strategic planning. The Modena ELLIS Unit was represented by Lorenzo Baraldi.
The retreat opened with short presentations in which each participant introduced their group and research lines, followed by open discussion sessions on the future directions of the program. The scientific agenda covered a broad range of topics in multimodal artificial intelligence, including egocentric vision, foundation models, medical imaging and document understanding. For the Modena Unit, the meeting was an occasion to present AImageLab’s work on multimodal large language models and retrieval-augmented generation, and to discuss new collaborations within the program.
Beyond the scientific sessions, the remote setting of the valley, reachable only by rack railway, offered plenty of opportunities to strengthen the connections within the program and the collaborative spirit of the ELLIS community.
The meeting was organised and hosted by the ELLIS Unit Barcelona and the Computer Vision Center (CVC), with the support of Barcelona City Council and the Secretariat of Digital Policies of the Catalan Government.
👉 Read the ELLIS news
👉 Learn more about the ELLIS Program “Multimodal Learning Systems”
