From August 24 to August 27, 2023, the 17th China Conference on Knowledge Graph and Semantic Computing (CCKS 2023), with the theme of "Knowledge Graph Empowers Artificial General Intelligence", was successfully held in Shenyang. Well-known scholars from Chinese and foreign universities such as Tsinghua University, Chinese Academy of Sciences, Fudan University and Tongji University, as well as R&D teamsin the industry, attended the event to share development trends and practical experience in related fields and promote industry-university-research cooperation. GTCOM, as a strategic partner of the event, demonstrated to participants its cross-language and multi-domain knowledge graph platform and GeWu big model for sharing their application and prospects in fields such as global defense security, smart city development, research data analysis and social media.

GTCOM participated as CCKS strategic partner

Atthe industrial forum, Dr. Chen Ziyan, Deputy Director of GTCOM 2030 Artificial Intelligence Research Institute, was invited to give a speech entitled "The Application of Personalized Knowledge and Data Fusion Large Models in Social Media".According to Chen, social media has been widely used by individuals and organizations, to convey cultural values, give policy guidelines, and guide public opinions on current affairs. At the same time, the number of social robots is increasing rapidly, making how to effectively utilize social media a hot topic. Against this background, Chen introduced GTCOM’s exploration of key technologies such as AI content identification, personalized content generation, and anthropomorphic behavior generation using the company’s GeWu big model technology that integrates knowledge and data. He also shared several cases of big model technology application in personalized social media.

GeWu is a big model system that GTCOM released in November 2022. It includes a multi-language pre-training model, a multi-language machine translation ultra-large model, a multi-modal pre-training model and a multi-language generative dialogue model. At present, the generative big model has been used in application scenarios such as end-to-end generation of multi-language knowledge graphs, social media behavior generation, and report writing. In the future, focusing on social media, the company will conduct in-depth research along the technical route of ultra-large-scale generative models similar to ChatGPT, targeting at multi-modal anthropomorphic generation, empathy, and a metaverse that blends virtual and physical reality, contributing to the long-term development of artificial intelligence.