Generating Chinese Poetry from Images via Concrete and Abstract Information
This work addresses the challenge of generating semantically consistent poetry from images for applications in creative AI, though it is incremental as it builds on existing methods for image-to-poetry generation.
The paper tackles the problem of topic drift and semantic inconsistency in generating classical Chinese poetry from images by extracting and integrating concrete and abstract information from images, resulting in poems with better consistency without losing quality, as shown by automatic and human evaluations.
In recent years, the automatic generation of classical Chinese poetry has made great progress. Besides focusing on improving the quality of the generated poetry, there is a new topic about generating poetry from an image. However, the existing methods for this topic still have the problem of topic drift and semantic inconsistency, and the image-poem pairs dataset is hard to be built when training these models. In this paper, we extract and integrate the Concrete and Abstract information from images to address those issues. We proposed an infilling-based Chinese poetry generation model which can infill the Concrete keywords into each line of poems in an explicit way, and an abstract information embedding to integrate the Abstract information into generated poems. In addition, we use non-parallel data during training and construct separate image datasets and poem datasets to train the different components in our framework. Both automatic and human evaluation results show that our approach can generate poems which have better consistency with images without losing the quality.