|
|
A chatbot for the banking domain
Schmidtová, Patrícia ; Dušek, Ondřej (advisor) ; Rosa, Rudolf (referee)
This thesis designs, implements and evaluates a task-based chatbot, which is expected to answer questions and give advice from the banking domain. We present an extendable natural language understanding (NLU) module based on GATE Framework which serves to create interpretations of the user's utterance. We implement a rule-based dialog manager component which is responsible for answering based on the NLU's interpretations and a stored context. We also implement a template-based natural language generation module. We then evaluate the chatbot with human testers, verifying it performs well in most cases and identifying areas for future improvement. 1
|
|
|
GTTG - application for manipulation with train timetable diagrams
Sýkora, Jakub ; Ježek, Pavel (advisor) ; Rosa, Rudolf (referee)
Applications for organization of rail transport often contain visualization of traffic situations, referred to as train timetable diagrams. The goal of this thesis is to create a library that simplifies the development of such applications by providing graphical component which allows developers to implement their own custom train timetable dia- grams. The component, developed for .NET platform, utilizes 2D graphics library SkiaSharp for drawing and can be integrated into various GUI frameworks. The library features interactive rendered content of train timetable diagrams - for example, zooming in on specific areas of the content or clicking on visualized elements. The library also offers an easily extendable implementation of the basic train timetable diagram. As an example of our library utilization, we also developed a WPF application inten- ded for viewing train timetable diagrams of Czech railways. 1
|
|
|
Tvorba závislostního korpusu pro jorubštinu s využitím paralelních dat
Oluokun, Adedayo ; Zeman, Daniel (advisor) ; Rosa, Rudolf (referee)
The goal of this thesis is to create a dependency treebank for Yorùbá, a language with very little pre-existing machine-readable resources. The treebank follows the Universal Dependencies (UD) annotation standard, certain language-specific guidelines for Yorùbá were specified. Known techniques for porting resources from resource-rich languages were tested, in particular projection of annotation across parallel bilingual data. Manual annotation is not the main focus of this thesis; nevertheless, a small portion of the data was verified manually in order to evaluate the annotation quality. Also, a model was trained on the manual annotation using UDPipe.
|
|
|
Discovering the structure of natural language sentences by semi-supervised methods
Rosa, Rudolf ; Žabokrtský, Zdeněk (advisor) ; Tiedemann, Jörg (referee) ; Horák, Aleš (referee)
Discovering the structure of natural language sentences by semi-supervised methods Rudolf Rosa In this thesis, we focus on the problem of automatically syntactically ana- lyzing a language for which there is no syntactically annotated training data. We explore several methods for cross-lingual transfer of syntactic as well as morphological annotation, ultimately based on utilization of bilingual or multi- lingual sentence-aligned corpora and machine translation approaches. We pay particular attention to automatic estimation of the appropriateness of a source language for the analysis of a given target language, devising a novel measure based on the similarity of part-of-speech sequences frequent in the languages. The effectiveness of the presented methods has been confirmed by experiments conducted both by us as well as independently by other respectable researchers. 1
|
|
|
Coreference resolution for Universal Dependencies
Faryad, Ján ; Novák, Michal (advisor) ; Rosa, Rudolf (referee)
Title: Coreference resolution for Universal Dependencies Author: Ján Faryad Department: Institute of Formal and Applied Linguistics Supervisor: Mgr. Michal Novák Abstract: Coreference is an important tool for maintaining of the text coherence. Up to now, there has been no possibility to mark it in Universal Dependencies (UD), which is a project for universal description of morphology and dependency syntax. This work presents a way how to mark coreference in the UD project. It also includes a conversion of data with coreference annotation from the corpora PDT 3.0 and OntoNotes 5.0 with using a tool UDPipe for an automatic analysis of text in the UD style. This work is also aimed to implement a system for automatic resolution of pronoun coreference using machine learning. Finally, the quality of the system is evaluated by simple way. The design of the program emphasizes the language independence and compatibility with the Udapi interface, which is used for processing of the UD data. Keywords: coreference resolution, coreference, anaphora, Universal Dependencies, UD
|
|
|
Spojité reprezentace vět v neuronovém strojovém překladu
Cífka, Ondřej ; Bojar, Ondřej (advisor) ; Rosa, Rudolf (referee)
Recent advances in natural language processing using neural networks have given rise to numerous methods of obtaining continuous-space vector representations of textual data that can be exploited for various applications. One of these methods is to use internal representations learned by neural machine translation (NMT) models. However, the attention mechanism in modern NMT systems removes the single point in the neural network from which the source sentence representation can be extracted. In this thesis, we propose and empirically evaluate novel ways to remove this limitation. We review existing methods of obtaining sentence representations and evaluating them, and present novel intrinsic evaluation metrics. Next, we describe our modifications to attention-based NMT architectures that allow extracting sentence representations. In the experimental section, we analyze these representations and evaluate them using a wide range of metrics with a focus on meaning representation. The results suggest that the better the translation quality, the worse the performance on these tasks. We also observe no performance gains from using multi-task training to control the representations.
|
|
|
Deciphering tool for puzzlehunt games
Hlásek, Filip ; Mareček, David (advisor) ; Rosa, Rudolf (referee)
The thesis focuses on substitution ciphers used at puzzlehunt games. First, we collect samples of the language which is used in the texts, that we are interested in. Furthermore, we propose a language model specially designed for working with sparse data. After that, we will explore the ways of searching for probable solutions and we will present a straightforward algorithm. Then we will improve it and make it more efficient. A significant part of the project is a console application, which serves as a deciphering tool. It is able to solve more than 15 % out of the ciphers, on which it has been tested. The result can be improved if the user inputs his current geography location. In this case, the program will search just for nearby places. That will allow the deciphering tool to explore more options and achieve a better precision. Powered by TCPDF (www.tcpdf.org)
|
|
|
Automatic post-editing of phrase-based machine translation outputs
Rosa, Rudolf ; Mareček, David (advisor) ; Žabokrtský, Zdeněk (referee)
We present Depfix, a system for automatic post-editing of phrase-based English-to-Czech machine trans- lation outputs, based on linguistic knowledge. First, we analyzed the types of errors that a typical machine translation system makes. Then, we created a set of rules and a statistical component that correct errors that are common or serious and can have a potential to be corrected by our approach. We use a range of natural language processing tools to provide us with analyses of the input sentences. Moreover, we reimple- mented the dependency parser and adapted it in several ways to parsing of statistical machine translation outputs. We performed both automatic and manual evaluations which confirmed that our system improves the quality of the translations.
|
|
|
Web Interface for the Treex Framework
Sedlák, Michal ; Popel, Martin (advisor) ; Rosa, Rudolf (referee)
This work deals with a web application called Treex::Web which serves as a web interface for NLP framework Treex. The work addresses several Treex issues (e.g. absence of graphical user interface and complicated installation) and offers Treex::Web as a possible solution. At the beginning of this work we introduce the Treex framework itself. The following chapters describe Treex::Web's user interface (chapter 3) and the implementation of the whole web application (chapter 4). Conclusion of this work includes a comparison of NLP frameworks similar to Treex and their web interfaces. 1
|
|
|
OprDU - Software for Correcting Homework
Rosa, Rudolf ; Knap, Tomáš (advisor) ; Kulič, Luboš (referee)
It has been discovered that there is no suitable tool for teachers which would make correcting of homework easier for them. Therefore the goal of this work is to create a complex yet flexible software solution for students and teachers for dealing with written homework. The solution combines advantages of current approaches to the issue and eliminates their disadvantages. By its components it covers all phases of work on the document, i.e. its creation by the student, submission to the teacher, correction by the teacher, sending the corrected homework to the student and displaying the corrections to the student. The main part of the solution is an application for the teacher which enables him to make the corrections. It contains advanced functions, e.g. spell-checking or document export. There is an emphasis on maximum effectiveness and user-friendliness. An XML format has been created to represent the documents, which supports not only easy marking of corrections but insertion of other information as well. A server communicating via Web services has been created for a more sophisticated management of the documents and for communication between teachers and students. The last component is a simple student application for documents handling.
|
guest ::
