Keyphrase Generation: A Text Summarization Struggle
This work addresses the challenge of generating valuable keyphrases not present in the text for scientific articles, but it is incremental as it shows no improvement over existing methods.
The paper tackled the problem of generating keyphrases for scientific articles by treating them as abstractive summaries of titles and abstracts, but found that advanced text summarization neural architectures did not outperform simpler unsupervised or existing supervised methods in evaluations on four test datasets.
Authors' keyphrases assigned to scientific articles are essential for recognizing content and topic aspects. Most of the proposed supervised and unsupervised methods for keyphrase generation are unable to produce terms that are valuable but do not appear in the text. In this paper, we explore the possibility of considering the keyphrase string as an abstractive summary of the title and the abstract. First, we collect, process and release a large dataset of scientific paper metadata that contains 2.2 million records. Then we experiment with popular text summarization neural architectures. Despite using advanced deep learning models, large quantities of data and many days of computation, our systematic evaluation on four test datasets reveals that the explored text summarization methods could not produce better keyphrases than the simpler unsupervised methods, or the existing supervised ones.