New achievement of DeepO Band: Composing Persian Modern poetry by Artificial Intelligence Using Language Modeling and Deep Learning
Mohammad Hasan Olyaei Torqabeh1*, Ali Olyaei Torqabeh2 and Hosein Olyaei Torqabeh3
1. Faculty of Electrical Engineering, Sadjad University of Technology, Mashhad, Iran,
*این آدرس ایمیل توسط spambots حفاظت می شود. برای دیدن شما نیاز به جاوا اسکریپت دارید
2. Department of Computer Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran, این آدرس ایمیل توسط spambots حفاظت می شود. برای دیدن شما نیاز به جاوا اسکریپت دارید
3. Department of Mechanical Engineering, Faculty of Engineering, Mashhad Branch, Islamic Azad University, Mashhad, Iran, این آدرس ایمیل توسط spambots حفاظت می شود. برای دیدن شما نیاز به جاوا اسکریپت دارید
Abstract
Today Language modeling has made many advances by scholars, and in the areas of speech processing, translation of texts has found many uses. Language modeling is a tool for learning the details and rules of a language to artificial intelligence. In this paper, first the original idea of the DeepO Band is described in order to compose new Persian poems. Then the recurrent neural network (RNN) and its details and operation are described and the algorithm for learning Persian poetry is explained. Simulation and coding of this paper are done using the Python language. The collection of modern poems for training and modeling are 300 lines of modern poetry, which are about 6,000 words, has been selected from the collection of poems by Sohrab Sepehri and contemporary poets. Finally, to test the trained model, five modern poems have been written by artificial intelligence. Simulation results show that modeling of modern poems is well done by artificial intelligence, and only 3.8% of this poems are similar to trained poems, and 96.2% of the rest are different poems.
Keywords: Language Modeling, Artificial Intelligence, Modern Poems, DeepO Band, Recurrent Neural Network, python
