China launches intelligent language processing tool for ancient books

BEIJING, Dec. 18 (Xinhua) -- Confused by obscure ancient expressions? An intelligent ancient language processing tool may give a hand.

China has launched its first intelligent tool "Xunzi," designed specifically for the processing and research of ancient books.

It is a large language model of ancient books, including the "Siku Quanshu," or the "Complete Library in the Four Branches of Literature," and collects more than two billion words.

Developed by researchers from the College of Information Management of Nanjing Agricultural University, the large language model can carry out functions like natural language understanding, automatic translation, poetry generation and automatic indexing.

The model can facilitate common users to understand the ancient expressions, punctuate and translate into modern Chinese language. Thanks to this smart tool, professionals can conduct lexical analysis, text classification, matching, and abstract generation.

The language model is available as an open-source tool on both GitHub and ModelScope websites, according to the researchers.

Xunzi was a renowned Chinese philosopher during the Warring States Period (475 B.C.-221 B.C.) and his namesake book is a compilation of his philosophical writings. The large language model was named "Xunzi" in honor of the philosopher.

The researchers said they look forward to cultivating more interdisciplinary talents through the intelligent study of ancient books and continuing to inherit the Chinese cultural heritage.

The language tool is also expected to provide convenience to more ancient book readers and promote the creative development of Chinese traditional culture.

The "Xunzi" large language model will be also applied in artificial intelligence (AI) writing, AI teaching and digital entertainment.