Keyphrase Generation with Correlation Constraints
This work solves keyphrase generation for text summarization and information retrieval, but it is incremental as it builds on existing sequence-to-sequence approaches with specific enhancements.
The paper tackled the problem of automatic keyphrase generation by addressing duplication and coverage issues in conventional methods, proposing CorrRNN, a sequence-to-sequence architecture that uses coverage vectors and preceding phrases to capture correlations, resulting in significant outperformance of state-of-the-art methods on benchmark datasets in accuracy and diversity.
In this paper, we study automatic keyphrase generation. Although conventional approaches to this task show promising results, they neglect correlation among keyphrases, resulting in duplication and coverage issues. To solve these problems, we propose a new sequence-to-sequence architecture for keyphrase generation named CorrRNN, which captures correlation among multiple keyphrases in two ways. First, we employ a coverage vector to indicate whether the word in the source document has been summarized by previous phrases to improve the coverage for keyphrases. Second, preceding phrases are taken into account to eliminate duplicate phrases and improve result coherence. Experiment results show that our model significantly outperforms the state-of-the-art method on benchmark datasets in terms of both accuracy and diversity.