ROetry: First Steps Towards Computer Generated Poetry in Romanian Language
Autor: | Mihai-Lica Pura, Elena-Casiana Fusu |
---|---|
Rok vydání: | 2020 |
Předmět: |
Poetry
Rhyme Computer science Romanian media_common.quotation_subject 05 social sciences Sign (semiotics) Linguistics language.human_language Field (computer science) Test (assessment) Recurrent neural network language State (computer science) 0509 other social sciences 050904 information & library sciences media_common |
Zdroj: | 2020 12th International Conference on Electronics, Computers and Artificial Intelligence (ECAI). |
DOI: | 10.1109/ecai50035.2020.9223175 |
Popis: | Over the past decades, the scientists in the field of technology dreamt of a utopia in which computers are human-like machines, capable of comprehending emotions and feelings. Poetry, the sign of culture and intelligence, could be the final test to determine if computers are capable of expressing themselves. Can a computer write poetry? This is the question whose answer could make art and science, the two spheres of knowledge that have seemed to always exclude each other, finally meet.This paper approaches the state of automatic poetry generation by studying the current achievements and the methodologies developed so far for this purpose. We propose a Romanian-based application developed by using both Markov chain and LSTM recurrent neural network models. The goal is to generate a text with metrical constraints (rhyme, number of syllables), that can be read as poetry. |
Databáze: | OpenAIRE |
Externí odkaz: |