以多模态大模型推动中国古文字研究发展

Promoting the Development of Ancient Chinese Character Research with Multimodal Large Models

人工智能技术与古文字研究的结合已经产生了一批有影响的成果,但几乎都是基于单模态模型。单一模态的智能模型在面对复杂的古文字问题时还存在很大的限制,只能在某些方面起到“辅助”作用,难以独立而完整地解决各种实际问题。文字的形、音、义与各种模态存在不同的关联,汉字的表意属性决定了图像模态和文本模态具有同样重要的作用,中国古文字研究与人工智能的结合须走多模态之路。近年来多模态大模型正在飞速发展,以此为契机来综合性地解决古文字检测、识别、复原等问题,有望取得更好的效果,而文字考释等一些过去难以触及的关键问题也能探索出新的解决路径。

The combination of artificial intelligence technology and ancient Chinese characters research has produced several influential results,but almost all of them are single-modal.The single-modal intelligence model has great limitations in facing the complex problems of ancient characters,and it can only play an“auxiliary”role in some aspects,and it is difficult to solve various practical problems independently and completely.The shape,sound and meaning of characters have different associations with various modalities,and the ideographic properties of Chinese characters determine that the image modality and the text modality have the same important roles,so it is more important to take the multimodal path for the combination of the research of ancient Chinese characters and artificial intelligence.In recent years,the multimodal large model is developing rapidly,which is an opportunity to comprehensively solve the problems of recognition,detection and restoration of ancient Chinese characters,which is expected to achieve better results,and some key problems such as textual interpretation,which were difficult to reach in the past,can also be explored to find a new solution path.