We study approaches to improve fine-grained short answer Question Answering models by integrating coarse-grained data annotated for paragraph-level relevance and show that coarsely annotated data can bring significant performance gains. Experiments demonstrate that the standard multi-task learning approach of sharing representations is not the most effective way to leverage coarse-grained annotations. Instead, we can explicitly model the latent fine-grained short answer variables and optimize the marginal log-likelihood directly or use a newly proposed \emph{posterior distillation} learning objective. Since these latent-variable methods have explicit access to the relationship between the fine and coarse tasks, they result in significantly larger improvements from coarse supervision.
@article{arxiv.1811.02076,
title = {Improving Span-based Question Answering Systems with Coarsely Labeled Data},
author = {Hao Cheng and Ming-Wei Chang and Kenton Lee and Ankur Parikh and Michael Collins and Kristina Toutanova},
journal= {arXiv preprint arXiv:1811.02076},
year = {2018}
}