English

Large-scale representation learning from visually grounded untranscribed speech

Computer Vision and Pattern Recognition 2019-09-20 v1 Computation and Language Sound Audio and Speech Processing

Abstract

Systems that can associate images with their spoken audio captions are an important step towards visually grounded language learning. We describe a scalable method to automatically generate diverse audio for image captioning datasets. This supports pretraining deep networks for encoding both audio and images, which we do via a dual encoder that learns to align latent representations from both modalities. We show that a masked margin softmax loss for such models is superior to the standard triplet loss. We fine-tune these models on the Flickr8k Audio Captions Corpus and obtain state-of-the-art results---improving recall in the top 10 from 29.6% to 49.5%. We also obtain human ratings on retrieval outputs to better assess the impact of incidentally matching image-caption pairs that were not associated in the data, finding that automatic evaluation substantially underestimates the quality of the retrieved results.

Keywords

Cite

@article{arxiv.1909.08782,
  title  = {Large-scale representation learning from visually grounded untranscribed speech},
  author = {Gabriel Ilharco and Yuan Zhang and Jason Baldridge},
  journal= {arXiv preprint arXiv:1909.08782},
  year   = {2019}
}
R2 v1 2026-06-23T11:19:51.361Z