A Hierarchical Model for Data-to-Text Generation
Computation and Language
2019-12-23 v1 Information Retrieval
Machine Learning
Abstract
Transcribing structured data into natural language descriptions has emerged as a challenging task, referred to as "data-to-text". These structures generally regroup multiple elements, as well as their attributes. Most attempts rely on translation encoder-decoder methods which linearize elements into a sequence. This however loses most of the structure contained in the data. In this work, we propose to overpass this limitation with a hierarchical model that encodes the data-structure at the element-level and the structure level. Evaluations on RotoWire show the effectiveness of our model w.r.t. qualitative and quantitative metrics.
Cite
@article{arxiv.1912.10011,
title = {A Hierarchical Model for Data-to-Text Generation},
author = {Clément Rebuffel and Laure Soulier and Geoffrey Scoutheeten and Patrick Gallinari},
journal= {arXiv preprint arXiv:1912.10011},
year = {2019}
}
Comments
Accepted at the 42nd European Conference on IR Research, ECIR 2020