EXPLORA

EXPLORA-SCALE aims to advance scalable, sustainable AI through provable performance. Key research includes learning in low-data regimes (zero-shot, continual), theoretical analysis of models like CLIP, and robust adaptation via model merging and low-rank techniques. It explores synthetic data utility, efficient Transformer/GNN architectures, and integrating symbolic reasoning into DL. Led by UB with CVC, UA, and UJI, the project delivers principled frameworks for evolving AI environments.

Members involved

Research lines