On the morning of March 14, the University of South China held the launching ceremony for the upgrade of the "USC AI Retrieval + USC AI Librarian" system and officially released the new-generation AI academic engines, namely USC AI Retrieval and USC AI Librarian.

The newly upgraded system deeply integrates the technologies of domestic large models like DeepSeek and Kimi, breaks through the boundaries of traditional literature retrieval, and realizes a leapfrog upgrade from "information query" to "intelligent generation". The USC AI Retrieval, an AI academic engine with domestic large models at its core, provides full-chain support for the academic research of teachers and students. The system covers over 1.2 billion literature resources around the world. Users only need to input keywords or natural language descriptions, such as "Research Progress in the Ethics of Brain-Computer Interfaces in the Past Five Years", and a structured Chinese and English literature review can be generated within 3 seconds. It automatically marks the source journals, impact factors and links to the original texts, increasing the reading efficiency by 200%. Its "Precise Retrieval" mode can intelligently filter out 80% of redundant information, while the "Knowledge Graph" function dynamically displays research trends, core institutions and disciplinary correlations, helping users quickly locate high-value literature.

The USC AI Librarian "Nan Xiao Tu" launched simultaneously supports dual-modal interactions of voice and text, can answer inquiries about library services in real time, accurately recommend books, journals and database resources, and handle services such as borrowing inquiries and electronic document delivery online. In addition, the USC AI Librarian integrates more than 20 tools including AI translation, academic video analysis and research topic comparison, truly realizing the "7×24-hour scientific research external brain" service.