English

Latent Alignment of Procedural Concepts in Multimodal Recipes

Computation and Language 2021-01-14 v1 Artificial Intelligence

Abstract

We propose a novel alignment mechanism to deal with procedural reasoning on a newly released multimodal QA dataset, named RecipeQA. Our model is solving the textual cloze task which is a reading comprehension on a recipe containing images and instructions. We exploit the power of attention networks, cross-modal representations, and a latent alignment space between instructions and candidate answers to solve the problem. We introduce constrained max-pooling which refines the max-pooling operation on the alignment matrix to impose disjoint constraints among the outputs of the model. Our evaluation result indicates a 19\% improvement over the baselines.

Keywords

Cite

@article{arxiv.2101.04727,
  title  = {Latent Alignment of Procedural Concepts in Multimodal Recipes},
  author = {Hossein Rajaby Faghihi and Roshanak Mirzaee and Sudarshan Paliwal and Parisa Kordjamshidi},
  journal= {arXiv preprint arXiv:2101.04727},
  year   = {2021}
}

Comments

Published in ALVR 2020, a workshop in ACL 2020