SummaRuNNer: A Recurrent Neural Network based Sequence Model for Extractive Summarization of Documents

14 Nov 2016  ·  Ramesh Nallapati, FeiFei Zhai, Bo-Wen Zhou ·

We present SummaRuNNer, a Recurrent Neural Network (RNN) based sequence model for extractive summarization of documents and show that it achieves performance better than or comparable to state-of-the-art. Our model has the additional advantage of being very interpretable, since it allows visualization of its predictions broken up by abstract features such as information content, salience and novelty. Another novel contribution of our work is abstractive training of our extractive model that can train on human generated reference summaries alone, eliminating the need for sentence-level extractive labels.

PDF Abstract


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Text Summarization CNN / Daily Mail (Anonymized) SummaRuNNer ROUGE-1 39.6 # 8
ROUGE-2 16.2 # 7
ROUGE-L 35.3 # 11
Text Summarization CNN / Daily Mail (Anonymized) Lead-3 baseline ROUGE-1 39.2 # 9
ROUGE-2 15.7 # 10
ROUGE-L 35.5 # 10


No methods listed for this paper. Add relevant methods here