This paper presents a neural network model for Arabic Question Answering(QA), that is capable of generating answers to simple factoid questions, with the help of the facts in a knowledge-base. The model is constructed for sequence-to-sequence learning. It has the ability to consult the knowledge-base, and a corpus of question-answer pairs with their associated triples in the knowledge-base. The experimental results demonstrate that the proposed model can effectively handle the variations of questions and answers, and produce correct answers in natural language text by utilizing the facts stored in the knowledge-base.
Question Answering, Neural Network, Information Retrieval
