AUTOMATIC GENERATION OF POETRY
Organization
Arvutiteaduse Instituut
Abstract
The thesis will explore generative deep learning techniques for automatically generating poetry.
If the thesis is taken at the bachelor level, the student will explore known generative deep
learning techniques for the Estonian language. If the thesis is done at the master level, the work will focus on the exploration of novel deep learning generative methods. Moreover, the master student's study will test the algorithm for two languages (e.g., Estonian and English, English and French, etc.). In both cases, the student will collect a corpus of poetry for training and will integrate an evaluative framework to assess the goodness of the automatic procedure. The student is free to choose the type of poetry to generate: the genre, the literary movement (e.g., symbolism), or the preferred poet (Marie Under, William Shakespeare, Charles Baudelaire), etc.
If the thesis is taken at the bachelor level, the student will explore known generative deep
learning techniques for the Estonian language. If the thesis is done at the master level, the work will focus on the exploration of novel deep learning generative methods. Moreover, the master student's study will test the algorithm for two languages (e.g., Estonian and English, English and French, etc.). In both cases, the student will collect a corpus of poetry for training and will integrate an evaluative framework to assess the goodness of the automatic procedure. The student is free to choose the type of poetry to generate: the genre, the literary movement (e.g., symbolism), or the preferred poet (Marie Under, William Shakespeare, Charles Baudelaire), etc.
Graduation Theses defence year
2019-2020
Supervisor
Eduard Barbu
Spoken language (s)
Requirements for candidates
Level
Masters
Keywords
Application of contact
Name
Eduard Barbu
Phone
E-mail
Full Document