We present FieldGuideRAG, a training-free GraphRAG method in which a vision-language model, or VLM, recognizes fine-grained species by comparing the query wi…
机构:新加坡国立大学
来源:arXiv 2610.40971 | AI4Papers 论文推荐平台