Attend to Medical Ontologies: Content Selection for Clinical Abstractive\n Summarization

Sequence-to-sequence (seq2seq) network is a well-established model for text\nsummarization task. It can learn to produce readable content; however, it falls\nshort in effectively identifying key regions of the source. In this paper, we\napproach the content selection problem for clinical abstractive summarization\nby augmenting salient ontological terms into the summarizer. Our experiments on\ntwo publicly available clinical data sets (107,372 reports of MIMIC-CXR, and\n3,366 reports of OpenI) show that our model statistically significantly boosts\nstate-of-the-art results in terms of Rouge metrics (with improvements: 2.9%\nRG-1, 2.5% RG-2, 1.9% RG-L), in the healthcare domain where any range of\nimprovement impacts patients' welfare.\n