Location : Home > News > News & Events
News & Events
The Symposium on " Large-Scale AI Models" of China Institute for Social-Legal Studies of Shanghai Jiao Tong University Successfully Held
2024-03-18 from:上海交通大学凯原法学院 preview:

The Symposium on " Large-Scale AI Models" of China Institute for Social-Legal Studies of Shanghai Jiao Tong University Successfully Held


On the afternoon of January 16, the symposium on " large-scale AI models" hosted by the China institute for social-legal studies of Shanghai Jiao Tong University was successfully held in Room 206 on the fourth floor of the north campus of Xuhui Campus. The symposium focused on " large-scale AI models in the financial industry", hosted by Ji Weidong, senior professor of liberal arts at Shanghai Jiao Tong University and dean of the China Institute of Law and Society, Zhang Quanshi, associate professor of John Hopcroft Computer Science Center of Shanghai Jiao Tong University, Liu Zhiyi, chief scientist of Oriental Fortune AI Research Institute, Fu Xiaomin, former big data expert of Capital One and current CTO of Haomai Wealth, Zhang Ji, chairman of Shanghai Strategy Consulting, Zhou Xiaoyong, partner of Huma Investment, Zhang Kai, senior researcher of Oriental Wealth,Zheng Yejie, Senior Scientific Research Cooperation Manager of Oriental Fortune, Xu Yue, director of the administrative office of the China Law and Society Research Institute of Shanghai Jiao Tong University, and Zhao Zerui, assistant director of the AI Governance and Law Research Center of Shanghai Jiao Tong University, attended the discussion.


Dean Ji Weidong first welcomed the visit of the experts, and invited Dean Liu Zhiyi to introduce the basic situation and cooperation intention of the Oriental Fortune AI Research Institute. Dean Liu Zhiyi gave a brief introduction to the recent achievements of Oriental Fortune in the research and development of large-scale AI models. He said that he hopes to strengthen cooperation with universities to jointly establish a research ecology of financial artificial intelligence, focusing on the research of large-scale AI models in financial scenarios.


Professor Zhang introduced the past cooperation experience and provided three research directions on the value alignment and evaluation of large-scale AI models: (1) the logical rationality evaluation of large model decision-making/assistance; (2) the white-box of the black box model; (3) Evaluation system to improve the training efficiency of large-scale AI models. Professor Zhang Kai and Professor Zhang Quanshi conducted further technical discussions and exchanges.

Based on the existing discussions, Dean Ji Weidong put forward three ideas from the perspective of theory and governance: first, the limit and significance of the interaction and assistance of the financial model; the second is the data governance of the financial model; The third is the communication problem of complex systems in the context of large-scale AI models. Dean Liu Zhiyi and Professor Zhang Quanshi respectively discussed the relationship between large models and small models, the scale limit of large models, and the pricing mechanism of data transactions.

At the meeting, a number of practical experts put forward their views and suggestions on issues related to the governance of large models. Fu Xiaomin put forward suggestions on the application and landing scenarios of financial large-scale AI models, and recognized the research significance of data pricing and rights confirmation. Based on his personal investment experience, Zhang pointed out that the financial model has practical significance, and hopes to become an observer and user of the financial model. Zhou Xiaoyong talked more about the significance of the financial model for the intelligent transformation and angel investment of small enterprises from the micro perspective of corporate finance, and agreed with the advanced research and layout of Oriental Wealth.

In addition, the experts and scholars also conducted in-depth discussions on future cooperation and exchanges.

预览: {$content}