A Study on Multilingual Intelligent Dissemination and Adaptive Reconstruction of Cultural Identity of Hezhe Oral History Using Multilingual Large Model Few-shot Learning

Abstract

The Hezhe people, one of China’s smaller ethnic groups, preserve oral histories that record hunting and fishing culture, historical memory, and the basis of ethnic identity. In an era of globalization and digitalization, the dissemination of Hezhe oral history faces challenges such as limited language carriers, narrow audience reach, and weakened cultural imagery across contexts. For antenna-enabled digital museums, wireless cultural communication platforms, and electromagnetic-media display systems, reliable multilingual semantic encoding is also necessary to preserve cultural meaning during transmission. This study builds a multilingual intelligent dissemination framework for Hezhe oral history based on multilingual large models and Few-shot Learning. Using a parallel corpus of Hezhe, Chinese, English, and Russian, a prompt-based Few-shot translation and generation model is designed. Combined with a cultural specificity classification system and adaptive rewriting strategies, the framework supports high-quality conversion and dissemination of oral history content in multilingual environments. The study further verifies the effectiveness of the framework through quantitative evaluation and qualitative analysis across four dimensions: language quality, cultural fidelity, audience acceptance, and cultural identity impact. The results demonstrate that Few-shot Learning has advantages in low-resource language processing and that multilingual dissemination can support adaptive reconstruction of Hezhe cultural identity, providing a reusable technical path for digital preservation and wireless cultural communication.