A Large-Scale Multi-Length Headline Corpus for Analyzing Length-Constrained Headline Generation Model Evaluation
This work addresses a gap in evaluation for headline generation in news production, but it is incremental as it focuses on dataset creation and validation rather than new methods.
The authors tackled the problem of evaluating length-constrained headline generation models by introducing two Japanese corpora, JNC and JAMUL, and found that using the longest reference summary for evaluation has issues despite being common practice.
Browsing news articles on multiple devices is now possible. The lengths of news article headlines have precise upper bounds, dictated by the size of the display of the relevant device or interface. Therefore, controlling the length of headlines is essential when applying the task of headline generation to news production. However, because there is no corpus of headlines of multiple lengths for a given article, previous research on controlling output length in headline generation has not discussed whether the system outputs could be adequately evaluated without multiple references of different lengths. In this paper, we introduce two corpora, which are Japanese News Corpus (JNC) and JApanese MUlti-Length Headline Corpus (JAMUL), to confirm the validity of previous evaluation settings. The JNC provides common supervision data for headline generation. The JAMUL is a large-scale evaluation dataset for headlines of three different lengths composed by professional editors. We report new findings on these corpora; for example, although the longest length reference summary can appropriately evaluate the existing methods controlling output length, this evaluation setting has several problems.