Focus-Based Text Summarisation with Hybrid Embeddings

Text summarisation has been recognised as an important Natural Language Processing task, attracting great attention from both researchers and practitioners. It has been widely adopted in various domains. For example, text summarisation of news, articles and book chapters can produce a short text, assisting the readers with grasping the main idea rapidly. In the medical domain, it is also applied to summarise the patients' questions. However, it is very challenging to control the summariser output by producing domain-specific summaries since the focus of domain-specific information may be ignored. In this paper, we propose a novel summarisation model aiming at producing summaries by focusing on the domain-specific knowledge, where hybrid embeddings, i.e., focus, domain and context embeddings, are utilised. We conduct extensive experiments to evaluate our novel model by using the MeQSum dataset. The experimental results demonstrate that our model outperforms state-of-the-art algorithms.