To achieve state-of-the-art performance, keyphrase extraction systems rely on domain-specific knowledge and sophisticated features. In this paper, we propose a neural network architecture based on a Bidirectional Long Short-Term Memory Recurrent Neural Network that is able to detect the main topics on the input documents without the need of defining new hand-crafted features. A preliminary experimental evaluation on the well-known INSPEC dataset confirms the e↵ectiveness of the proposed solution.

This paper evaluates different techniques for building a supervised, multilanguage keyphrase extraction pipeline for languages which lack a gold standard. Starting from an unsupervised English keyphrase extraction pipeline, we implement pipelines for Arabic, Italian, Portuguese, and Romanian, and we build test collections for languages which lack one. Then, we add a Machine Learning module trained on a well-known English language corpus and we evaluate the performance not only over English but on the other languages as well. Finally, we repeat the same evaluation after training the pipeline over an Arabic language corpus to check whether using a language-specific corpus brings a further improvement in performance. On the five languages we analyzed, results show an improvement in performance when using a machine learning algorithm, even if such algorithm is not trained and tested on the same language.