RANLP 2019 Multilingual Headline Generation Task Overview

John M Conroy

Abstract: The objective of the 2019 RANLP Mul- tilingual Headline Generation (HG) Task is to explore some of the challenges high- lighted by current state of the art ap- proaches on creating informative head- lines to news articles: non-descriptive headlines, out-of-domain training data, generating headlines from long documents which are not well represented by the head heuristic, and dealing with multilingual domain. This tasks makes available a large set of training data for headline genera- tion and provides an evaluation methods for the task. Our data sets are drawn from Wikinews as well as Wikipedia. Partici- pants were required to generate headlines for at least 3 languages, which were eval- uated via automatic methods. A key as- pect of the task is multilinguality. The task measures the performance of multi- lingual headline generation systems using the Wikipedia and Wikinews articles in multiple languages. The objective is to as- sess the performance of automatic head- line generation techniques on text docu- ments covering a diverse range of lan- guages and topics outside the news do- main.