Towards Unsupervised Image Captioning with Shared Multimodal Embeddings

TL;DR

This paper proposes a novel unsupervised image captioning method using shared multimodal embeddings structured by visual concepts, enabling caption generation without explicit image-caption pairs.

Contribution

It introduces a shared latent space for images and text, trained with weak supervision and adversarial learning, allowing effective unsupervised image captioning.

Findings

Outperforms previous unsupervised methods

Learns semantically meaningful representations

Utilizes large text corpora for training

Abstract

Understanding images without explicit supervision has become an important problem in computer vision. In this paper, we address image captioning by generating language descriptions of scenes without learning from annotated pairs of images and their captions. The core component of our approach is a shared latent space that is structured by visual concepts. In this space, the two modalities should be indistinguishable. A language model is first trained to encode sentences into semantically structured embeddings. Image features that are translated into this embedding space can be decoded into descriptions through the same language model, similarly to sentence embeddings. This translation is learned from weakly paired images and text using a loss robust to noisy assignments and a conditional adversarial component. Our approach allows to exploit large text corpora outside the annotated distributions of image/caption data. Our experiments show that the proposed domain alignment learns a semantically meaningful representation which outperforms previous work.