Students taught the neural network to imitate Mayakovsky

The research team of students of the School of Digital Economy of the Far Eastern Federal University has developed artificial intelligence that can compose poetry. The neural network creates stepped poetic lines in the style of Vladimir Mayakovsky.

The development is built on the basis of a generative and aspect network-an algorithm of machine learning, which uses two neural networks, one of which generates samples, and the other tries to distinguish the correct samples from the wrong ones. For the training of neural networks, the FFFU underwear needed a week. During this time, artificial intelligence analyzed the works of Vladimir Mayakovsky, and then learned to imitate the poet's style.

“Machine training and artificial intelligence technology has great potential in digital creativity,” explained Ilya Mirin, director of the DVFU digital economy. - An art that creates a neural network can be integrated into partner projects. Students did a huge job, teaching artificial intelligence to creativity. They have a reserve for further success and even commercial use. ”

Earlier, students of the Far Eastern Federal University created a digital musician - the neural network performs works in the lounge style, which play in one of the coastal offices of Sberbank.