VALSE Webinar 20230614-13期 总第313期 面向医疗领域的基础大模型探索与应用 ... ...

报告嘉宾：刘天明 (University of Georgia)

报告题目：When Brain-Inspired AI Meets AGI

报告嘉宾：付杰 (北京智源人工智能研究院)

报告题目：Cross-Lingual Multi-Modal Language Models for Healthcare

Panel嘉宾：

刘天明 (University of Georgia)、付杰 (北京智源人工智能研究院)、李响 (哈佛大学医学院、麻省总医院)、许言午 (华南理工大学)、檀韬 (Macao Polytechnic University)、雷柏英 (深圳大学)、陈浩 (香港科技大学)

Panel议题：

1. ChatGPT在NLP对话方面取得了巨大的成功，医学健康领域有哪些未来可能爆发的研究与应用？

2. SAM (segment anaything)在视觉语义分割方面取得了很好的泛化性能，医疗视觉方面应该如何构建类似的基础大模型？

3. 多模态数据的融合与分析是未来人工智能医疗领域一个重点方向，未来在这个方向上应该怎样构建基础大模型？

4. 基础大模型如何更加高效适配到医疗领域的下游任务？

5. 如何看待基础大模型在医疗健康领域所涉及的安全伦理问题？例如如何实现大模型的可控生成等？

6. 基础大模型的训练需要强大的数据和硬件资源支撑，学术界如何应对这些挑战？挑战与机遇往往并存，那么医疗领域又有哪些新的机遇？

*欢迎大家在下方留言提出主题相关问题，主持人和panel嘉宾会从中选择若干热度高的问题加入panel议题！

报告嘉宾：刘天明 (University of Georgia)

报告时间：2023年06月14日 (星期三)晚上20:00 (北京时间)

报告题目：When Brain-Inspired AI Meets AGI

报告人简介：

Dr. Tianming Liu is a Distinguished Research Professor and a Full Professor of Computer Science at The University of Georgia. Dr. Liu’s research interests are brain imaging, computational neuroscience, brain-inspired artificial intelligence, and artificial general intelligence. Dr. Liu published 400+ research papers on these topics, his Google citation is over 11,000+, and his H-index is 55. Dr. Liu is the recipient of NIH Career Award and NSF CAREER Award. Dr. Liu serves on the editorial boards of multiple international journals including Medical Image Analysis, IEEE Transactions on Medical Imaging, IEEE Reviews in Biomedical Engineering, IEEE/ ACM Transactions on Computational Biology and Bioinformatics, and IEEE Journal of Biomedical and Health Informatics. Dr. Liu is a Fellow of AIMBE (American Institute of Medical and Biological Engineering)and was the General Chair of MICCAI 2019.

个人主页：

https://cobweb.cs.uga.edu/~tliu/

报告摘要：

ChatGPT and GPT-4 have demonstrated remarkable performance in many natural language processing (NLP), reasoning, content generation, and multimodal tasks, particularly, in zero-shot learning settings. In some sense, ChatGPT/ GPT-4 already exhibits human intelligence traits, or artificial general intelligence (AGI). In this talk, I will share my understanding of core technologies in ChatGPT/ GPT-4 including foundation models, large language models (LLM), in-context learning, prompt engineering, reinforcement learning from human feedback, and multimodal integration, and in particular, I will offer brain science perspectives to these innovative methodologies. Our recent research papers on brain-inspired AI and AGI will be discussed as examples.  

参考文献：

[1] Lin Zhao, Lu Zhang, Zihao Wu, Yuzhong Chen, Haixing Dai, Xiaowei Yu, Zhengliang Liu, Tuo Zhang, Xintao Hu, Xi Jiang, Xiang Li, Dajiang Zhu, Dinggang Shen, Tianming Liu, When Brain-inspired AI Meets AGI 

arxiv: https://arxiv.org/abs/2303.15935. 2023.

[2] Chong Ma, Zihao Wu, Jiaqi Wang, Shaochen Xu, Yaonai Wei, Zhengliang Liu, Lei Guo, Xiaoyan Cai, Shu Zhang, Tuo Zhang, Dajiang Zhu, Dinggang Shen, Tianming Liu, Xiang Li, ImpressionGPT: An Iterative Optimizing Framework for Radiology Report Summarization with ChatGPT, 

arxiv: https://arxiv.org/abs/2304.08448. 2023.

[3] Zhenxiang Xiao, Yuzhong Chen, Lu Zhang, Junjie Yao, Zihao Wu, Xiaowei Yu, Yi Pan, Lin Zhao, Chong Ma, Xinyu Liu, Wei Liu, Xiang Li, Yixuan Yuan, Dinggang Shen, Dajiang Zhu, Tianming Liu, Xi Jiang, Instruction-ViT: Multi-Modal Prompts for Instruction Learning in ViT, arxiv: https://arxiv.org/abs/2305.00201. 2023.

报告嘉宾：付杰 (北京智源人工智能研究院)

报告时间：2023年06月14日 (星期三)晚上20:30 (北京时间)

报告题目：Cross-Lingual Multi-Modal Language Models for Healthcare

报告人简介：

Jie Fu is a researcher at Beijing Academy of Artificial Intelligence. He received the Ph.D. degree from National University of Singapore, under the supervision of Tat-Seng Chua. He worked as a postdoc fellow with Yoshua Bengio at Quebec AI Institute (Mila), funded by Microsoft Research Montreal. He was an IVADO postdoc fellow working with Chris Pal at Quebec AI Institute (Mila). His current research interests include deep learning and large language model with their applications in NLP, CV, and other real-world tasks. He received ICLR 2021 outstanding paper award.

个人主页：

https://bigaidream.github.io/

报告摘要：

Large language models have made great progress in natural language processing (NLP). This inspires researchers to apply LLMs for fields that were not considered the core playground of NLP, for example, healthcare. However, the first bottleneck for medicine using LLMs is the data, like most other breakthroughs in artificial intelligence that starts with data collection. I will first talk about our efforts in data collection and discuss frameworks we design for such tasks.

参考文献：

[1] Zhongwei Wan, Che Liu, Mi Zhang, Jie Fu, Benyou Wang, Sibo Cheng, Lei Ma, César Quilodrán-Casas, Rossella Arcucci, “Med-UniC: Unifying Cross-Lingual Medical Vision-Language Pre-Training by Diminishing Bias,” arXiv:2305.19894 (2023).

[2] Jianquan Li, Xidong Wang, Xiangbo Wu, Zhiyi Zhang, Xiaolong Xu, Jie Fu, Prayag Tiwari, Xiang Wan, Benyou Wang, "Huatuo-26M, a Large-scale Chinese Medical QA Dataset," arXiv:2305.01526 (2023).

[3] Ge Zhang, Yemin Shi, Ruibo Liu, Ruibin Yuan, Yizhi Li, Siwei Dong, Yu Shu, Zhaoqun Li, Zekun Wang, Chenghua Lin, Wenhao Huang, Jie Fu, “Chinese Open Instruction Generalist: A Preliminary Release," arXiv:2304.07987 (2023).

[4] Zekun Wang, Ge Zhang, Kexin Yang, Ning Shi, Wangchunshu Zhou, Shaochun Hao, Guangzheng Xiong, Yizhi Li, Mong Yuan Sim, Xiuying Chen, Qingqing Zhu, Zhenzhu Yang, Adam Nik, Qi Liu, Chenghua Lin, Shi Wang, Ruibo Liu, Wenhu Chen, Ke Xu, Dayiheng Liu, Yike Guo, Jie Fu, “Interactive Natural Language Processing," arXiv:2305.13246 (2023).

Panel嘉宾：李响 (哈佛大学医学院、麻省总医院)

嘉宾简介：

李响博士毕业于上海交通大学自动化系本科，获美国佐治亚大学计算机科学博士，师从佐治亚大学杰出教授，美国医学与生物工程院 (AIMBE)会士刘天明教授。2016年毕业后任哈佛大学医学院和麻省总医院博士后，接受美国医学院院士，前麻省总医院放射科主任James Thrall和麻省总医院先进医学计算和分析中心 (CAMCA)主任李全政教授的指导，现任职讲师。李响博士主要从事机器学习，人工智能及大数据科学在医学健康领域中的解决方案研究和算法开发。李响博士在领域顶级期刊及国际会议上发表了百余篇学术论文，并于2019年起创办了International Workshop on Multiscale Multimodal Medical Imaging会议。其多项研究获得来自美国国立卫生研究院 (NIH)和麻省总医院的资助，并获得多项期刊与会议奖项。

个人主页：

https://xiangli-shaun.github.io/

Panel嘉宾：许言午 (华南理工大学)

嘉宾简介：

许言午博士，华南理工大学教授，前百度智慧医疗科学家，WHO数字健康咨询委员会专家，新加坡眼科研究所客聘教授，中国科学院宁波工研院客聘研究员，南方科技大学业界导师，IEEE高级会员，中国生物医学工程学会科技创新与产业促进工作委员会委员，中国医药教育协会数字影像与智能医疗专委会副主任委员，中国医药教育协会智能眼科分会常委、数字疗法工作委员会常委，北京卫生法学会大数据与互联网人工智能医疗专委会委员。自2004年起，他持续从事计算机视觉、机器学习理论及其应用研究，共发表了150余篇国际期刊及会议论文，谷歌引用6700余次，申请国际专利20多件和中国专利70多件。他目前担任医疗影像顶会MICCAI和IPMI组委，Springer Nature旗下Medical Imaging和BioMedical Engineering Online期刊编委，Frontiers in Public Health和Diagnostics期刊编委，中华医学会主办“中国科技期刊卓越行动计划”英文期刊Intelligent Medicine创刊编委，AAAI、ACPR、ACCAS等国际学术会议组委及PC委员，眼科医学影像国际会议OMIA和国际比赛平台iChallenge创始主席。他先后获聘公安部引智计划特聘专家、浙江省特聘专家、北京市特聘专家，获评MICCAI优秀审稿人、中国科协优秀审稿人、MICCAI卓越领域主席。他主持了超过2000万人民币的国家级科研项目和超过4000万人民币的横向课题，作为项目负责人成功入围工信部和国家药监局合办的第一次“AI医疗器械创新任务揭榜”。从0到1负责百度眼底AI产品获批全国首张多病种AI三类证 (同时也是国内首张青光眼AI三类证)。

个人主页：

https://sites.google.com/site/xuyanwu1982/home

Panel嘉宾：檀韬 (Macao Polytechnic University)

嘉宾简介：

Dr. Tan Tao is currently an associate professor at Macao Polytechnic University. He was a senior scientist at GE healthcare, responsible for enterprise AI projects. He was a guest assistant professor at Eindhoven University of Technology in the Netherlands. He is mainly engaged in the frontier research of artificial intelligence in medical imaging, and has done in-depth research in breast imaging. He has made contributions to the industry-university-research and scientific research transformation of medical artificial intelligence. He has published over 100 papers in his research field, including 10 articles with an IF of 10 or above. He has Led 4 FDA-approved products in the United States, 2 EU CE-approved products, and 1 Chinese NMPA-approved product, contributing over 20 million US dollars in economic output. He has 16 US patents pending. He received the GE Healthcare Innovation Award in 2020. 

个人主页：

https://scholar.google.com/citations?user=lLg3WRkAAAAJ&hl=en

Panel嘉宾：雷柏英 (深圳大学)

嘉宾简介：

雷柏英，国家级青年人才入选者，深圳大学和深圳大学总医院特聘教授，博士生导师，深圳市海外高层次人才 (孔雀计划)、深圳市高层次后备级人才，深圳市孔雀团队核心成员等，获新加坡南洋理工大学博士学位。先后在美国北卡大学教堂山分校和法国计算和自动化研究所等研究机构进行研究和访问。主要研究方向为医学图像处理和人工智能。在IEEE TMI、IEEE TNNLS、Medical Image Analysis 以第一/ 通讯作者 (含共同)发表SCI论文100多篇。谷歌学术总引用超7000次，H指数41。主持国家自然科学基金联合基金重点1项，面上2项等项目20余项 (含国家级7项)。现任IEEE TNNLS、IEEE TCYB、IEEE TMI、Medical Image Analysis、IEEE JBHI 等10种SCI期刊编委。IEEE高级会员，IEEE Bio Imaging Signal Processing (BISP)Technical Committee (TC)技术委员会委员，Biomedical Imaging and Image Processing (BIIP)TC技术委员会委员，医学图像顶级学术会议MICCAI 2021-2022领域主席。IEEE Guangzhou Section, Women in Engineering Affinity Group 主席，人工智能A类会议AAAI、IJCAI程序委员会委员；中国图象图形学会 (CSIG)，中国人工智能学会 (CAAI)高级会员。获吴文俊人工智能科学技术奖三等奖，深圳市科学技术奖一等奖。入选美国斯坦福大学发布的“全球前2%顶尖科学家” (2020-2022)，全球顶尖前10万科学家 (2021)，获2022“强国青年科学家”提名 (全国共40人)，CSIG石青云女科学家奖。

个人主页：

http://bme.szu.edu.cn/20181/0612/66.html

主持人：陈浩 (香港科技大学)

主持人简介：

陈浩，香港科技大学计算机科学与工程系和化学与生物工程系助理教授，IEEE高级会员，研究兴趣包括人工智能，医疗图像分析，深度学习等。他领导的人工智能医疗实验室 (Smart Lab)，专注于可信赖人工智能技术在医疗领域的研究与应用。陈博士于2017年获得香港中文大学博士学位，曾任香港中文大学博士后研究员及荷兰乌特勒支大学医学中心访问学者。在MICCAI、IEEE-TMI、MIA、CVPR、AAAI、IJCAI、Radiology、Lancet Digital Health、Nature Machine Intelligence、JAMA等顶级期刊和会议发表论文100余篇 (谷歌学术引用次数达到19500余次，h-index 60)，入选2022年斯坦福大学全球排名前2%科学家名单。此外，陈博士还具有丰富的工业研究和产业转化经验，拥有十余项人工智能和图像分析方面专利。曾获得2023年亚洲青年科学家、2019年人工智能医学影像顶级会议MICCAI青年科学家影响力奖、Elsevier-MICCAI 最佳论文奖、医学影像与增强现实会议最佳论文奖、福布斯中国30岁以下30位精英、香港资讯及通讯科技银奖等奖项，担任包括IEEE TNNLS、J-BHI、CMIG和Medical Physics等期刊编委，担任MICCAI 2021-2023、ISBI 2022、MIDL 2022-2023、AAAI 2022 等多个人工智能与医学影像分析国际会议的领域主席和程序委员，曾带领团队获得15余项国际医学图像分析的挑战赛冠军。2020年相关研究被Nature Index报道为改变人工智能未来应用的六名青年科学家之一。

个人主页：

https://cse.hkust.edu.hk/~jhc/

特别鸣谢本次Webinar主要组织者：

主办AC：陈浩 (香港科技大学)

协办AC：雷柏英 (深圳大学)