|
|
Morph classifier
John, Vojtěch ; Žabokrtský, Zdeněk (advisor) ; Helcl, Jindřich (referee)
Morphological classification is the task of classifying morphs - the forms of morphemes - in laready segmented words. Since there are more and greateer resources for morpholog- ical segmentation than for morphological classification, methods presented in this thesis could be used for enriching already existing resources or creating new ones. We propose several methods of morph classification in increasing order of granularity. Firstly, we present several unsupervised, semi-supervised and supervised methods of root identifica- tion (on eight languages with manually annotated data in UniSegments), using simple statistical methods, derivational tree databases and CNN-LSTM-CRF sequence classifi- cation. We sample the learning curve of the neural taggers and train joint models for morphological segmentation and classificaton. Further, we present supervised feature- independent morph classification and joint classification and segmentation CNN-LSTM- CRF models, trained and tested on Czech, Slovak and Russian. Finally, we proposed two simple methods of aligning morphological features to morphs. We also used interlin- ear glossed text databases to assign morphological functions to inflectional morphs using CNN-LSTM-CRF sequence classifiers. 1
|
|
|
English poetry generation with pretrained language models
John, Vojtěch ; Dušek, Ondřej (advisor) ; Musil, Tomáš (referee)
Title: English poetry generation with pretrained language models Author: Vojtěch John Department: Institute of Formal and Applied Linguistics Supervisor: Mgr. et Mgr. Ondřej Dušek, Ph.D., Institute of Formal and Applied Linguistics Abstract: In this thesis we deal with possible ways of English poetry generation independent on a concrete form of poetic form. We briefly discuss selected known attempts to English poetry generation with respect to their adaptability. Based on the already existing limerick generator LimGen we created a new English poetry generator called PGen, flexible with respect to rhyme and metric schemes and allowing usage of finetuned language models. We present PGen together with subjective evaluation of poems it generated (in 4 different poetic forms) by several human raters. Keywords: poetry generation, natural language generation, natural language processing, pretrained language models
|
guest ::
RSS feed.