AI RESEARCH
Prune Redundancy, Preserve Essence: Vision Token Compression in VLMs via Synergistic Importance-Diversity
arXiv CS.CV
•
ArXi:2603.09480v1 Announce Type: new Vision-language models (VLMs) face significant computational inefficiencies caused by excessive generation of visual tokens. While prior work shows that a large fraction of visual tokens are redundant, existing compression methods struggle to balance importance preservation and information diversity. To address this, we propose PruneSID, a