Key words: didactics - German as a foreign language - communicative competence - free writing Abstract: The aim of the presented master's thesis is to point out the role of free writing in German as a foreign language. The thesis is divided into two parts: theoretical and practical. The theoretical part summarizes elemental knowledge of writing skills, its classification and evaluation. Special attention is given to the concept of free writing. The practical part compares theoretical knowledge with realized research. The research analyses the attitude of the target group to free writing and its correction.

The purpose of this thesis is to map out the beginning of literature written in the domestic language, i.e. Galician-Portuguese and later Portuguese, as it started to separate from works written in Latin. The first part of the thesis describes how the Portuguese language evolved from Latin, which in turn evolved from the Indo-European language. A colloquial form of Latin began to emerge, splitting later into several dialects. The Galician- Portuguese language appears and individual Romance languages start to form. The new Portuguese language undergoes a period of formation, with the first phase being the Old Portuguese, followed by the classical period, until Portuguese became the language we know now. The second part of the thesis describes the formation and early phases of Portuguese writing, both literary and non-literary, including a list of genres that prevailed during the beginnings of Portuguese literature, and a selection of excerpts of the oldest texts and their translations. Key words: Old Portuguese, old texts, evolution of language, Portuguese philology, translation

The presented Master's thesis entitled "Pronounciation mistakes of Czech students learning Portuguese" tries to depict the most common mistakes in the pronounciation of Portuguese sounds by Czech students. The thesis is divided into two parts: theoretical and practical. The theoretical part, based on a careful research of specialized materials, deals with phonetics, the phonological aspect of a language, phonology, the description of speech forming processes, sound systems of Portuguese, rules for correct pronounciation of respective sounds and also interference - the influences of a person's mother tongue while learning Portuguese. The theoretical part served as a basis for the practical part. In the practical part an analysis was carried out, using a number of audio recordings taken from first-year to fifth-year students of Portuguese Studies at Charles University, Faculty of Arts. The aim of this analysis was to detect typical mistakes made by Czech Students in the pronounciation of Portuguese sounds. The observed pronounciation phenomena are recorded in tables, each phenomena separately. In conclusion, the thesis provides an evaluation of the analysis.

The thesis compares methods and forms of Czech language teaching at secondary grammar school and Waldorf lycée. The first part deals with the theoretical basis of traditional and alternative education, whereas the second part focuses on their practical implementation as methods and forms of work. The observed lessons are analysed and the results of traditional and alternative approach are contrasted by a test. The aim is to ascertain whether the alternative school implements its principles into methods and forms of language teaching and whether they are more effective than the traditional ones. The thesis concludes that the alternative school implements only a few aspects of alternative education resulting in more active and engaged pupils. The results of the test imply that both schools are comparable in terms of language knowledge. KEYWORDS teaching Czech language, syntax, upper-secondary school, grammar school, waldorf lycée

The goal of this thesis is to explore the area of natural language correction and to design and implement neural network models for a range of tasks ranging from general grammar correction to the specific task of diacritization. The thesis opens with a description of existing approaches to natural language correction. Existing datasets are reviewed and two new datasets are introduced: a manually annotated dataset for grammatical error correction based on CzeSL (Czech as a Second Language) and an automatically created spelling correction dataset. The main part of the thesis then presents design and implementation of three models, and evaluates them on several natural language correction datasets. In comparison to existing statistical systems, the proposed models learn all knowledge from training data; therefore, they do not require an error model or a candidate generation mechanism to be manually set, neither they need any additional language information such as a part of speech tags. Our models significantly outperform existing systems on the diacritization task. Considering the spelling and basic grammar correction tasks for Czech, our models achieve the best results for two out of the three datasets. Finally, considering the general grammatical correction for English, our models achieve results which are...

Machine learning has been paid a lot of attention in recent years. One of the studied fields is employment of adversarial examples. These are artifi- cially constructed examples which evince two main features. They resemble the real training data and they deceive already trained model. The ad- versarial examples have been comprehensively investigated in the context of deep convolutional neural networks which process images. Nevertheless, their properties have been rarely examined in connection with NLP-processing networks. This thesis evaluates the effect of using the adversarial examples during the training of the recurrent neural networks. More specifically, the main focus is put on the recurrent networks whose text input is in the form of a sequence of word/character embeddings, which have not been pretrained in advance. The effects of the adversarial training are studied by evaluating multiple NLP datasets with various characteristics.

The purpose of this diploma thesis is to analyse the ability of the human ear to hear slight rhythm deviations in speech and non-speech phrases. The first part contains theoretical background for the study of speech rhythm and summarizes the research that has been already conducted in this area. It focuses especially on the perceptual nature of rhythm, the concept of P-centers, and provides a comparative study of speech rhythm and musical rhythm and their common properties and functions. The theoretical part is concluded with the analysis of potential influences of linguistic and musical training on the production and perception of rhythm, and hypotheses and research questions are formulated. The practical part contains a perceptual experiment designed to examine the ability to identify rhythm manipulations in short speech and non-speech, i.e. percussive, phrases. Short English phrases are selected and their rhythmically altered counterparts are prepared. Participants are then presented with pairs of speech or non-speech phrases and a task to identify rhythmical discrepancies between them. The results highlighted several differences between the nature of speech and non-speech rhythm. While the presence of stressed syllables enhances perception of rhythm deviations in speech, this is not the case...

This diploma thesis is study of development of eductional application with game features. Theme of the aplication is developing of social skills. Introduction consists of determination of requirements on process of development and on application itself. The work follows with theoretical basis for particular aspects of developing and selling of application. From technical perspective the work contains detail informations about so called chatbot technologies, it's princips of functioning and evaluation of it's contemporary maturity level with regard to usability in my application. The work ends by concrete proposal of solutin of application based on theoretical findings.

This dissertation addresses the problem of classification of entities in text represented by noun phrases. The goal of this thesis is to develop a method for automated classification of entities appearing in datasets consisting of short textual fragments. The emphasis is on unsupervised and semi-supervised methods that will allow for fine-grained character of the assigned classes and require no labeled instances for training. The set of target classes is either user-defined or determined automatically. Our initial attempt to address the entity classification problem is called Semantic Concept Mapping (SCM) algorithm. SCM maps the noun phrases representing the entities as well as the target classes to WordNet. Graph-based WordNet similarity measures are used to assign the closest class to the noun phrase. If a noun phrase does not match any WordNet concept, a Targeted Hypernym Discovery (THD) algorithm is executed. The THD algorithm extracts a hypernym from a Wikipedia article defining the noun phrase using lexico-syntactic patterns. This hypernym is then used to map the noun phrase to a WordNet synset, but it can also be perceived as the classification result by itself, resulting in an unsupervised classification system. SCM and THD algorithms were designed for English. While adaptation of these algorithms for other languages is conceivable, we decided to develop the Bag of Articles (BOA) algorithm, which is language agnostic as it is based on the statistical Rocchio classifier. Since this algorithm utilizes Wikipedia as a source of data for classification, it does not require any labeled training instances. WordNet is used in a novel way to compute term weights. It is also used as a positive term list and for lemmatization. A disambiguation algorithm utilizing global context is also proposed. We consider the BOA algorithm to be the main contribution of this dissertation. Experimental evaluation of the proposed algorithms is performed on the WordSim353 dataset, which is used for evaluation in the Word Similarity Computation (WSC) task, and on the Czech Traveler dataset, the latter being specifically designed for the purpose of our research. BOA performance on WordSim353 achieves Spearman correlation of 0.72 with human judgment, which is close to the 0.75 correlation for the ESA algorithm, to the author's knowledge the best performing algorithm for this gold-standard dataset, which does not require training data. The advantage of BOA over ESA is that it has smaller requirements on preprocessing of the Wikipedia data. While SCM underperforms on the WordSim353 dataset, it overtakes BOA on the Czech Traveler dataset, which was designed specifically for our entity classification problem. This discrepancy requires further investigation. In a standalone evaluation of THD on Czech Traveler dataset the algorithm returned a correct hypernym for 62% of entities.

The thesis focuses on the processing of morphologically complex words in Czech. It follows up on the previous research in the field of inflectional morphology, in which Czech is minimally represented. The thesis is divided into two parts. The first part describes the main theoretical approaches, which are connectionism and dualism. The second part describes an experiment, which was carried out following the previous research and aimed at comparing the claims of the two main approaches. The experiment was conducted on 37 native speakers of Czech aged from eight to eleven. The participants' task was to say as quickly as possible the genitive form of a noun that appeared on the screen of an electronic device in nominative form. The results of the research do not support the dualistic thoughts, but rather support the connectionist view. However, further empirical research is needed for a clear conclusion. Key words speeded production; inflected word form; frequency; language acquisition; inflection