Don’t Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization

• DOI: 10.18653/v1/d18-1206
• Publication Date: 2019-06-29T20:01:17Z
• Authors: Shashi Narayan; Shay B. Cohen; Mirella Lapata
• Keywords: FOS: Computer and information sciences; Computer Science - Computation and Language; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology; Computation and Language (cs.CL); 01 natural sciences; 0105 earth and related environmental sciences

We introduce “extreme summarization”, a new single-document summarization task which does not favor extractive strategies and calls for an abstractive modeling approach. The idea is to create short, one-sentence news summary answering the question “What article about?”. collect real-world, large-scale dataset this by harvesting online articles from British Broadcasting Corporation (BBC). propose novel model conditioned on article’s topics based entirely convolutional neural networks. demonstrate experimentally that architecture captures long-range dependencies in document recognizes pertinent content, outperforming oracle system state-of-the-art approaches when evaluated automatically humans.