Abstract:Features of designing a system of automated generation of poetic works, which opens up new opportunities for artistic speech and show business, especially the preparation of poems and songs have been considered. Quite often lyrics without special content become successful due to the lack of complex plots, as well as due to the unobtrusiveness and ease of perception by listeners. Well-known literature sources and available software products that can generate poetic works by combining different methods and algorithms are analyzed. It has been established that none of them is able to ensure the content and uniqueness of the poetic work at the same time, especially in the Ukrainian language. The existing approaches to the generation of poetic works are analysed, among which the relevant is a method based on templates, generation and testing, evolutionary algorithms and the method based on specific cases. Peculiarities of generating poetical works, first of all rhyming rules, types of strophes, poetic rhythms and sizes have been investigated. An approach to automated generation of poetic works using evolutionary algorithms and a method based on specific cases have been developed. Their combination resembles a sequence of actions for creative personalities when creating poems or writing lyrics. Peculiarities of neural network organization for automated generation of poetic works have been considered. It is proposed to perform neural network training using the method of inverse propagation and using a genetic algorithm. The principle of operation of algorithms for finding optimal solutions which contain such consecutive stages as initialization, evaluation of solutions, population selection, evolution of solutions, is analysed. Their interaction and various opportunities for neural network learning have been investigated in detail. An algorithm has been developed according to which the software application will analyse the poetic works offered by the user and generate new variants of it on the basis received from the neural network of logically connected words or lines of the verse in the poem. The user can edit both the components of the poem and the generated poetic works, and thus can train the neural network. The specification of requirements to the software application has been developed, the basic requirements to the user interface are defined, and also potential classes of users who will use it are established. Keywords: computational linguistics; Artificial Intelligence; neural network; genetic algorithm; optimal solution.

Shall I Compare Thee to a Machine-Written Sonnet? An Approach to Algorithmic Sonnet Generation

Poetry Generation Algorithm with Automatic Expansion of Keyword Semantic Information

Romantic-Computing

Introducing Aspects of Creativity in Automatic Poetry Generation

Composing Like an Ancient Chinese Poet: Learn to Generate Rhythmic Chinese Poetry

Zero-shot Sonnet Generation with Discourse-level Planning and Aesthetics Features

There Once Was a Really Bad Poet, It Was Automated but You Didn't Know It

Can Machine Generate Traditional Chinese Poetry? A Feigenbaum Test

GPoeT-2: A GPT-2 Based Poem Generator

Can AI Write Classical Chinese Poetry like Humans? An Empirical Study Inspired by Turing Test

Let's FACE it. Finnish Poetry Generation with Aesthetics and Framing

Erato: Automatizing Poetry Evaluation

Neural Poetry: Learning to Generate Poems using Syllables

Beyond Narrative Description: Generating Poetry from Images by Multi-Adversarial Training

Sonnet or Not, Bot? Poetry Evaluation for Large Models and Datasets

Post-Automation Poetics, or How Cold-War Computers Discovered Poetry

Deep-speare: A Joint Neural Model of Poetic Language, Meter and Rhyme

Augmenting Poetry Composition with Verse by Verse

Let the Poem Hit the Rhythm: Using a Byte-Based Transformer for Beat-Aligned Poetry Generation

Design of the system of automated generation of poetry works

Evaluating Diversity in Automatic Poetry Generation