Visual Storytelling via Predicting Anchor Word Embeddings in the Stories
This work addresses the problem of generating coherent narratives from image sequences for applications in AI-assisted content creation, though it appears incremental as it builds on existing seq2seq methods with a novel embedding prediction component.
The authors tackled visual storytelling by predicting anchor word embeddings from images and using them with image features to generate narrative sentences, achieving the best results in most automatic evaluation metrics and outperforming competing methods in human evaluation.
We propose a learning model for the task of visual storytelling. The main idea is to predict anchor word embeddings from the images and use the embeddings and the image features jointly to generate narrative sentences. We use the embeddings of randomly sampled nouns from the groundtruth stories as the target anchor word embeddings to learn the predictor. To narrate a sequence of images, we use the predicted anchor word embeddings and the image features as the joint input to a seq2seq model. As opposed to state-of-the-art methods, the proposed model is simple in design, easy to optimize, and attains the best results in most automatic evaluation metrics. In human evaluation, the method also outperforms competing methods.