基于深度学习的文化符号传播机制与受众认知偏差

Transmission Mechanism of Cultural Symbols and Audience Cognitive Bias Based on Deep Learning

摘要: 文化符号的数字化传播与内涵传承是当前文化数字化发展落地的核心命题，多模态深度学习技术为相关研究提供了新的技术路径。本文以敦煌飞天为典型研究样本，先后完成文化符号多模态表征体系搭建、深度学习驱动的传播链路模拟与扩散规律梳理、受众认知偏差分层分类与靶向修正框架搭建三类核心研究工作，明确了从特征编码到落地传播全流程的标准化操作路径。相关研究成果可为中华优秀传统文化符号的数字化传承、跨圈层传播与受众认知引导提供可复用的实操参考，降低文化符号传播过程中的内涵损耗风险。

Abstract: The digital communication and connotation inheritance of cultural symbols constitute a core issue in the implementation of the current digital development of culture. Multimodal deep learning technologies offer new technical approaches for relevant research. Taking Dunhuang Flying Apsaras as a typical research sample, this paper carries out three core research tasks sequentially: constructing a multimodal representation system for cultural symbols, simulating deep learning-driven communication links and summarizing diffusion rules, as well as establishing a framework for hierarchical classification and targeted correction of audience cognitive bias. A standardized operation workflow covering the whole process from feature encoding to practical communication is clarified. The relevant research findings can provide reusable practical references for the digital inheritance, cross-group communication and audience cognition guidance of fine traditional Chinese cultural symbols, and reduce the risk of connotation loss during the communication of cultural symbols.