Generalist Vision-Language Models for Fast Radio Burst detection: a zero-shot benchmark against a specialized detector
用于快速射电暴检测的通用视觉语言模型:针对专用探测器的零样本基准测试
机构 * Universidade Federal de Campina Grande(坎皮纳格兰德联邦大学) ; Instituto de Formação de Educadores, Universidade Federal do Cariri(卡里里联邦大学教育工作者培训学院)
专题命中 视觉定位与Grounding :vision-language model(title,abstract);VLM(abstract_cn);分类 cs.LG
AI总结 研究通用视觉语言模型在零样本下检测快速射电暴,从模拟动态频谱中抽取样本作基准,比较其与专用探测器,发现Gemma 4 2B准确率与SwinYNet相近,在结构化RFI上误报率低,重写提示可用于三类分类,准确率较高。
Comments 31 pages, 7 figures. Section added with analysis for real data. New figures and tables added. Other minor changes