YNU-HPCC at SemEval-2020 Task 10: Using a Multi-granularity Ordinal Classification of the BiLSTM Model for Emphasis Selection

Abstract

Written text emphasis in visual media is used to increase the comprehension of written text, to grab a viewer's attention, and to convey the author's intent.The task is choosing candidates for emphasis in short written text, to enable automated design assistance in authoring. As the author's intent is unknown and only the input text is available, multiple emphasis selections are valid. In this study, we propose a multi-granularity ordinal classification method to address the problem of emphasis selection. Specifically, word embeddings are learned from the Embeddings from Language Model (ELMo) to extract feature vector representations. Then, the ordinal classifications are implemented on four different multi-granularities to approximate the continuous emphasized values. Comparative experiments were conducted to compare the model with the baseline, in which the problem is transformed to a label distribution problem.The code of this paper is availabled at: https://github.com/DavidInWuhanChina/SemEval-2020-Task10.