Wentao Fan

Multimodal topic modeling aims to uncover interpretable semantic structures from documents containing both text and images, while also supporting downstream tasks such as multimodal retrieval. However, existing multimodal neural topic models, typically based on standard Variational Autoencoders (...