AI RESEARCH
HAC: Parameter-Efficient Hyperbolic Adaptation of CLIP for Zero-Shot VQA
arXiv CS.CV
•
ArXi:2604.23665v1 Announce Type: new Recent advances in representation learning have shown that hyperbolic geometry can offer a expressive alternative to the Euclidean embeddings used in CLIP models, capturing hierarchical structures and leading to better-organized representations. However, current hyperbolic CLIP variants are trained entirely from scratch, which is computationally expensive and resource-intensive.