Omni-Persona: Systematic Benchmarking and Improving Omnimodal Personalization
Omni-Persona:系统性基准测试与改进多模态个性化
机构 * Department of Electrical and Computer Engineering, Seoul National University(首尔国立大学电气与计算机工程系) ; Interdisciplinary Program in Artificial Intelligence, Seoul National University(首尔国立大学人工智能跨学科项目) ; Department of Artificial Intelligence, University of Seoul(首尔大学人工智能系)
专题命中 视觉定位与Grounding :grounding(abstract);multimodal large language model(abstract);分类 cs.CV
AI总结 Omni-Persona提出首个多模态个性化基准,通过Persona Modality Graph任务组和细粒度任务,系统分析多模态个性化中的接地行为,提出Calibrated Accuracy评估方法,揭示开源模型在音频与视觉接地上的差距及SFT与RLVR的局限性。
Comments Project Page: https://github.com/oyt9306/Omni-Persona