Query-oriented text summarization based on hypergraph transversals
作者:
Highlights:
• A theoretical connection is made between sentence retrieval and the extraction of a transversal in a hypergraph.
• A new computationally efficient topic model is proposed, based on the semantic clustering of terms.
• In our hypergraph model, each hyperedge represents a topic of the corpus, and each sentence is tagged with multiple topics.
• The proposed algorithms are computationally cheaper than existing hypergraph-based summarizers.
• On real-world datasets, our model outperforms existing graph-based summarization systems by 6% of ROUGE-SU4 F-measure.
摘要
•A theoretical connection is made between sentence retrieval and the extraction of a transversal in a hypergraph.•A new computationally efficient topic model is proposed, based on the semantic clustering of terms.•In our hypergraph model, each hyperedge represents a topic of the corpus, and each sentence is tagged with multiple topics.•The proposed algorithms are computationally cheaper than existing hypergraph-based summarizers.•On real-world datasets, our model outperforms existing graph-based summarization systems by 6% of ROUGE-SU4 F-measure.
论文关键词:Query-oriented text summarization,Hypergraph theory,Hypergraph transversal,Sentence clustering,Submodular set functions
论文评审过程:Received 10 January 2018, Revised 20 January 2019, Accepted 6 March 2019, Available online 27 March 2019, Version of Record 27 March 2019.
论文官网地址:https://doi.org/10.1016/j.ipm.2019.03.003