Author(s)

Ziyu Li

Corresponding Author

Ziyu Li

Abstract

The interaction between European literature and music has a rich and extensive history, spanning from ancient Greek epics and medieval troubadour poetry to Renaissance madrigals and Romantic German lieder. This study establishes a trend analysis system for the interaction between literature and music, utilizing a multimodal large language model. The core components of both the large language model and the audio analysis model are based on the Transformer architecture, which features key technologies such as the Self-Attention Mechanism and Positional Encoding. By learning from extensive textual and audio data, the model possesses cross-disciplinary and multilingual analytical capabilities. In this study, we analyzed ten musical works and their corresponding poetic texts from five different historical periods using the model. The model achieved an accuracy rate of 93.57% on 140 test questions. Based on the model's analysis, this study provides a summary of the historical trends in the interaction between European literature and music, focusing on musical elements, emotional tone, and socio-cultural aspects.

Keywords

Music, Literature, Art History, Large Language Model, Deep Learning, Audio-Text Multimodality