Deep Learning 4 Postmodern Poetry

Featured projects (1)

At least 80 per cent of modern and postmodern poems have neither rhyme nor metrical schemes such as iambic or trochaic meter. Does this, however, mean that they lack any rhythmical features? According to US research on free verse prosody, the opposite is true: modern poets like Whitman, the Imagists, the Beat poets, and contemporary Slam poets developed a post-metrical idea of prosody that employs rhythmical features of everyday language, prose, and musical styles including jazz and hip hop. Our project will test this theory by applying digital pattern recognition techniques to a corpus of modern and postmodern poems as read aloud by the original authors. To this end, we examine the four major online portals for spoken poetry. Making use of the vastly improved prosody detection available in speech processing technology today, we will identify rhythmical features through methods including phrase break prediction, prosodic phrasing, spoken document analysis, and fluency/disfluency modeling. In a first step, the philological sub-project will define rhythmical patterns based on a comparison of the textual line arrangement with the prosodic phrasing of the poet’s voice. Then the digital sub-project will develop an automatic pattern recognition tool, based on machine learning techniques, that is then capable of analyzing further material. Our aim is to develop a methodology and a software tool for prosody detection and formal corpus analysis, tailored to written and spoken modern free verse poetry. In traditional poetry, metrical patterns such as pentameter or hexameter are used to identify poetic forms (for instance the elegiac couplet) or influences (for instance the influence of ancient Greek poetry on eighteenth-century German writers). In a similar way, our study of rhythmical patterns will help to identify poetic forms in free verse prosody and detect influences such as that of American free verse on modern and postmodern German writers. In the long run, our tool will be hosted by the German website, and will be accessible for university teaching and research in a sustainable way.

Featured research (2)

The translation of poetry is a complex, multifaceted challenge: the translated text should communicate the same meaning, similar metaphoric expressions, and also match the style and prosody of the original poem. Research on machine poetry translation is existing since 2010, but for four reasons it is still rather insufficient: 1) The few approaches existing completely lack any knowledge about current developments in both lyric theory and translation theory. 2) They are based on very small datasets. 3) They mostly ignored the neural learning approach that superseded the long-standing dominance of phrase-based approaches within machine translation. 4) They have no concept concerning the pragmatic function of their research and the resulting tools. Our paper describes how to improve the existing research and technology for poetry translations in exactly these four points. With regards to 1) we will describe the ‘Poetics of Translation’. With regards to 2) we will introduce the Worlds largest corpus for poetry translations from lyrikline. With regards to 3) we will describe first steps towards a neural machine translation of poetry.With regards to 4) we will describe first steps towards the development of a poetry translation mapping system.

Lab head

Burkhard Meyer-Sickendiek
  • Institute of German and Dutch Languages and Literature

Members (12)

Markus Conrad
  • Universidad de La Laguna
Timo Baumann
  • University of Hamburg
Hussein Hussein
  • Freie Universität Berlin
Michael Dambacher
  • Universität Konstanz
Guillermo Recio
  • University of Barcelona
Annette Kinder
  • Freie Universität Berlin
David Schmidtke
  • Freie Universität Berlin
Isabel Bohrn
  • Freie Universität Berlin