Zhichao Han
Algorithm Specialist, eBay Payments Risk Control
He graduated from Shanghai Jiaotong University with a Bachelor's Degree in Information Security, and then obtained a Master's Degree in Computer Science from the University of Manchester. He worked as an algorithm engineer in the R&D departments of asset management and investment banks, focusing on technological innovation and practical application in the field of financial technology. In eBay Payment Risk Control Department, he led the development of experimental system based on graph mining technology and the AIGC workflow project integrating production and research, which effectively promoted the application of graph algorithms and deep learning models in the field of anti-fraud and compliance, and obtained a number of technology patents. He has published several academic papers in top international conferences such as KDD, AAAI, VLDB and ICDE, and is committed to transforming cutting-edge technologies into enterprise-level solutions to drive business innovation and development. Translated with DeepL.com (free version)
Topic
Development Strategies for Convergence of Large and Traditional Models in eBay E-Commerce Businesses
With the rapid development of AI technology, big models show great potential in theory, but also face challenges in practical applications such as generating illusions and high latency in reasoning. eBay adopts innovative big model and traditional model fusion development strategies in its core e-commerce business areas, such as recommender systems, anti-fraud mechanisms, and compliance management, effectively avoiding and solving these challenges to more efficiently utilise big model technology to drive business innovation and development. The company has been able to effectively avoid and solve these challenges, and thus more efficiently utilise big model technology to drive business innovation and development. In this seminar, we will discuss how eBay has optimised the synergy of models in these key business areas. We will also demonstrate how this strategy can accurately improve decision quality and operational efficiency, and share our valuable experience in realising model complementarity and enhancing business value.
