Relative Drawing Identification Complexity is Invariant to Modality in Vision-Language Models
机构 * Interactive Technologies Institute and NOVA LINCS Faculty of Exact Sciences and Engineering University of Madeira Portugal(互动技术研究所和NOVA LINCS精确科学与工程学院马德拉大学) ; Department of Informatics University of Bergen Norway(信息学院卑尔根大学挪威) ; Valencian Research Institute for Artificial Intelligence Universitat Politècnica de València Spain(瓦伦西亚人工智能研究机构瓦伦西亚理工大学西班牙) ; Leverhulme Centre for the Future of Intelligence and Valencian Research Institute for Artificial Intelligence Spain(未来智能中心和瓦伦西亚人工智能研究机构西班牙)
专题命中 其他VLM :vision-language model(title,abstract);分类 cs.CV
Comments 54 pages (42 pages of appendix). Accepted for publication at the ECAI 2025 conference