|
|
Low-Resource Neural Machine Translation
Filo, Denis ; Fajčík, Martin (referee) ; Jon, Josef (advisor)
This thesis deals with neural machine translation (NMT) for low-resource languages. The goal was to evaluate current techniques by using the experiments and suggest their improvements. The translation systems in this thesis used the neural network transformer architecture and were trained by the Marian framework. The selected language pairs were Slovak with Croatian and Slovak with Serbian. The subjects of the experiments were the transfer learning techniques and semi-supervised learning.
|
|
|
Migration and Regional Identity: The Case of Ukrainian Students in the Region of South Bohemia
GIBON, Gaëlle Reine Ingke
This master thesis on migration and regional identity focuses on the case of Ukrainian student migration in the region of South Bohemia in Czech Republic. It is an exploratory qualitative research. The research question of this study is: How do Ukrainian students perceive the regional identity of South Bohemia and how do they integrate in that region? The objectives behind that question were to see if these migrants perceive a regional identity in that region, if they identify with it and the main challenges they face to integrate.
|
|
|
Interplay Between Nomination Tendencies in the Language of Mass Media. The Case of the Analytical Adjectives
Vačkov, Veselin ; Gladkova, Hana (advisor) ; Rangelová, Albena (referee) ; Uhlířová, Ludmila (referee)
The thesis aims to describe and theoretically explain the range of complex nominal word structures and phrases that evolved in all Slavic languages with staggering productivity during the last decades. In particular, it focuses on the so called analytical adjectives, i.e. formally invariant lexical units mostly of foreign origin that manifest their attributive syntactic role (and the word class they belong to) not through a morpheme but through fixed word order position. The thesis rejects the view that analytical adjectives are first parts of compound words and explores several transitional cases. In more general terms, the thesis explores phenomena that test the boundary between words and phrases, morphology and syntax. It concludes that in Bulgarian there already exists a well established word subclass of analytical adjectives. They have resulted from an originally lexical innovation that has been grammatcalized due to intensive borrowings of words and phrase models mainly from English and their replication. The comparative study of present-day Bulgarian and Czech provides evidence of the strategies that the two genetically related languages use to achieve their nomination goals. The search for an explanation of the similarities and differences between both languages offers insights into their...
|
|
|
Low-Resource Neural Machine Translation
Filo, Denis ; Fajčík, Martin (referee) ; Jon, Josef (advisor)
This thesis deals with neural machine translation (NMT) for low-resource languages. The goal was to evaluate current techniques by using the experiments and suggest their improvements. The translation systems in this thesis used the neural network transformer architecture and were trained by the Marian framework. The selected language pairs were Slovak with Croatian and Slovak with Serbian. The subjects of the experiments were the transfer learning techniques and semi-supervised learning.
|
|
| |
|
| |
|
|
Linguistic Issues in Machine Translation between Czech and Russian
Klyueva, Natalia ; Kuboň, Vladislav (advisor) ; Panevová, Jarmila (referee) ; Strossa, Petr (referee)
In this thesis we analyze machine translation between Czech and Russian languages from the perspective of a linguist. We work with two types of Machine Translation systems - rule-based (TectoMT) and statistical (Moses). We experiment with different setups of these two systems in order to achieve the best possible quality. One of the questions we address in our work is whether relatedness of the discussed languages has some impact on machine translation. We explore the output of our two experimental systems and two commercial systems: PC Translator and Google Translate. We make a linguistically-motivated classification of errors for the language pair and describe each type of error in detail, analyzing whether it occurred due to some difference between Czech and Russian or is it caused by the system architecture. We then compare the usage of some specific linguistic phenomena in the two languages and state how the individual systems cope with mismatches. For some errors, we suggest ways to improve them and in several cases we implement those suggestions. In particular, we focus on one specific error type - surface valency. We research the mismatches between Czech and Russian valency, extract a lexicon of surface valency frames, incorporate the lexicon into the TectoMT translation pipeline and present...
|
|
|
Interplay Between Nomination Tendencies in the Language of Mass Media. The Case of the Analytical Adjectives
Vačkov, Veselin ; Gladkova, Hana (advisor) ; Rangelová, Albena (referee) ; Uhlířová, Ludmila (referee)
The thesis aims to describe and theoretically explain the range of complex nominal word structures and phrases that evolved in all Slavic languages with staggering productivity during the last decades. In particular, it focuses on the so called analytical adjectives, i.e. formally invariant lexical units mostly of foreign origin that manifest their attributive syntactic role (and the word class they belong to) not through a morpheme but through fixed word order position. The thesis rejects the view that analytical adjectives are first parts of compound words and explores several transitional cases. In more general terms, the thesis explores phenomena that test the boundary between words and phrases, morphology and syntax. It concludes that in Bulgarian there already exists a well established word subclass of analytical adjectives. They have resulted from an originally lexical innovation that has been grammatcalized due to intensive borrowings of words and phrase models mainly from English and their replication. The comparative study of present-day Bulgarian and Czech provides evidence of the strategies that the two genetically related languages use to achieve their nomination goals. The search for an explanation of the similarities and differences between both languages offers insights into their...
|
|
|
Morphonological features of Sanskrit and their Slavic parallels in the context of relationships between Indo-Iranian and Balto-Slavic languages
Džunková, Katarína ; Kostič, Svetislav (advisor) ; Vacek, Jaroslav (referee)
The present diploma thesis deals with the common morphological and phonological features between Sanskrit and Slavic languages. It contains the list of common lexemes in Sanskrit and Slavic languages added at the end of the thesis. The point of departure of this thesis are the theories of comparative linguistics, which are mentioned in the introduction. Morphophonological parallels are researched in the context of the relationships between Indo- Iranian and Balto-Slavic languages, what helps to distinguish common features between the separated language branches: e. g. common features between Baltic languages and Sanskrit or common features between Slavic and Iranian languages. Iranian and Slavic language contact is analysed in the special chapter. The common features between Sanskrit and Slavic languages supposed to be generally the remnants of common Indo-European principles, which are preserved in both language branches. The special chapter is also devoted to the RUKI sound law, which is regarded as one of the most important common feature between Balto-Slavic and Indo-Iranian languages.
|
|
|
On two expressions of negative mental state (hatred, fear)
Nejedlý, Petr
There is no indisputable recorded evidence of a fundamental non-negative form of the Old Church Slavic and Slavic verb nenaviděti (it is necessary to revise some opposite views in this respect). Later records of the verb naviděti can be, from the semantic point of view, explained rather as prefi xed derivatives of viděti. Nenaviděti had therefore in all likelihood originated as an artifi cially created denotation of an abstract concept based on the spheres of Christian doctrine. 2. As primarily Old Czech records indicate, the word family of the Slavic verb plašiti contains the IE root *pel-, the initial meaning of which is one of motion: 'chase, run aft er'. The sememe '(rapid/hasty) motion' needs to be considered as fundamental for already the earliest stages of Slavic languages; the sememe 'fright, fear' is – along with the corresponding intransitive verbs – secondary.
|
guest ::
