**支付宝发布多模态医疗大模型，携手20家机构推进AI医疗共建计划**
在近日举办的2024世界人工智能大会“可信大模型论坛”上，支付宝宣布了其在医疗AI领域的最新进展。该公司发布了多模态医疗大模型，并联合多家机构发起AI医疗共建计划，力图通过科技力量，使AI更深入地理解医疗行业。
经过半年多的测试打磨，支付宝医疗大模型正式亮相，成为国内首批多模态医疗大模型之一。蚂蚁集团大模型应用部总经理顾进杰详细介绍了背后的技术研发。蚂蚁百灵大模型具备多模态能力，可以直接理解和训练多模态数据。
支付宝发布的医疗大模型以此为基础，融入了包括报告、影像、药品等在内的百亿级中英文图文数据。该模型在医疗专业能力评测中表现卓越，特别是在中文医疗LLM评测榜单promptCBLUE中位列A榜第一，凸显了其高效和准确性。
此外，全栈自研的医疗多模态结构使得模型在识别报告、药品等图像方面的准确率超过90%。相关技术的先进性不仅在论坛中受到广泛赞誉，更在CVPR、KDD、ECCV等行业顶尖技术期刊与会议上发表。
在现场，支付宝还与包括人民卫生出版社、北京大学医学部等20家机构共同发起AI医疗共建计划。该计划旨在共同探索AI数智技术底座与创新应用服务，通过联合研发、资源共享等方式推动医疗AI的进步。这一合作无疑将为医疗行业带来革命性的变革，并为患者和医疗机构提供更高效、精准的医疗服务。
英语如下：
News Title: Alipay Launches Multimodal Medical Macro-model, Jointly Exploring the New Era of AI Medicine with 20 Organizations
Keywords: Alipay Medical Macro-model, AI Co-construction Plan, Multimodal Technology
News Content: **Alipay Unveils Multimodal Medical Macro-model, Joining Hands with 20 Organizations to Boost AI Medical Co-construction Plan**
At the “Trusted Macro-model Forum” of the World AI Conference 2024, Alipay announced its latest advancements in medical AI. The company launched a multimodal medical macro-model and teamed up with multiple organizations to initiate an AI medical co-construction plan, aiming to deepen AI’s understanding of the medical industry through technology.
After more than half a year of testing and polishing, the Alipay medical macro-model has officially been launched, becoming one of the first batch of multimodal medical macro-models in China. Gu Jinjie, the General Manager of Ant Group’s Macro-model Application Department, provided a detailed introduction to the technological research and development behind it. The Ant Baiding Macro-model possesses multimodal capabilities and can directly understand and train multimodal data.
The medical macro-model launched by Alipay is based on this and integrates billions of Chinese and English graphic data, including reports, images, and medicines. The model has demonstrated excellent performance in medical professional ability assessments, particularly ranking first in the Chinese medical LLM evaluation list promptCBLUE, highlighting its efficiency and accuracy.
In addition, the fully self-developed medical multimodal structure enables the model to achieve over 90% accuracy in recognizing images of reports, medicines, etc. The advanced technology not only received widespread praise at the forum but also was published in industry-leading technical journals and conferences such as CVPR, KDD, and ECCV.
At the scene, Alipay also jointly launched the AI medical co-construction plan with 20 organizations including People’s Health Publishing House and Beijing University of Medicine. The plan aims to jointly explore the technology base and innovative application services of AI digital technology, promoting the progress of medical AI through joint research and development and resource sharing. This collaboration will undoubtedly bring revolutionary changes to the medical industry and provide more efficient and accurate medical services for patients and medical institutions.
【来源】https://www.jiqizhixin.com/articles/2024-07-05-2
Views: 3