|
| |
|
|
Novel Methods for Natural Language Generation in Spoken Dialogue Systems
Dušek, Ondřej ; Jurčíček, Filip (advisor) ; Ircing, Pavel (referee) ; Žabokrtský, Zdeněk (referee)
Title: Novel Methods for Natural Language Generation in Spoken Dialogue Systems Author: Ondřej Dušek Department: Institute of Formal and Applied Linguistics Supervisor: Ing. Mgr. Filip Jurčíček, Ph.D., Institute of Formal and Applied Linguistics Abstract: This thesis explores novel approaches to natural language generation (NLG) in spoken dialogue systems (i.e., generating system responses to be presented the user), aiming at simplifying adaptivity of NLG in three respects: domain portability, language portability, and user-adaptive outputs. Our generators improve over state-of-the-art in all of them: First, our gen- erators, which are based on statistical methods (A* search with perceptron ranking and sequence-to-sequence recurrent neural network architectures), can be trained on data without fine-grained semantic alignments, thus simplifying the process of retraining the generator for a new domain in comparison to previous approaches. Second, we enhance the neural-network-based gener- ator so that it takes preceding dialogue context into account (i.e., user's way of speaking), thus producing user-adaptive outputs. Third, we evaluate sev- eral extensions to the neural-network-based generator designed for producing output in morphologically rich languages, showing improvements in Czech generation. In...
|
|
|
Unsupervised Dependency Parsing
Mareček, David ; Žabokrtský, Zdeněk (advisor) ; Jurčíček, Filip (referee) ; Sogaard, Anders (referee)
Unsupervised dependency parsing is an alternative approach to identifying relations between words in a sentence. It does not require any annotated treebank, it is independent of language theory and universal across languages. However, its main disadvantage is its so far quite low parsing quality. This thesis discusses some previous works and introduces a novel approach to unsupervised parsing. Our dependency model consists of four submodels: (i) edge model, which controls the distribution of governor-dependent pairs, (ii) fertility model, which controls the number of node's dependents, (iii) distance model, which controls the length of the dependency edges, and (iv) reducibility model. The reducibility model is based on a hypothesis that words that can be removed from a sentence without violating its grammaticality are leaves in the dependency tree. Induction of the dependency structures is done using Gibbs sampling method. We introduce a sampling algorithm that keeps the dependency trees projective, which is a very valuable constraint. In our experiments across 30 languages, we discuss the results of various settings of our models. Our method outperforms the previously reported results on a majority of the test languages.
|
|
|
Goal Oriented and Open Domain Dialogue Management
Vodolán, Miroslav ; Jurčíček, Filip (advisor) ; Psutka, Josef (referee) ; Šedivý, Jan (referee)
Title: Goal Oriented and Open Domain Dialogue Management Author: Miroslav Vodolán Department: Institute of Formal and Applied Linguistics Supervisor: Ing. Mgr. Filip Jurčíček, Ph.D., Institute of Formal and Applied Linguistics Abstract: This thesis proposes novel approaches for dialogue management in dialogue sys- tems. It covers goal-oriented and open domain dialogue systems. In both setups, it helps to improve quality of dialogues between the system and its users: 1) In the case of goal-oriented dialogues, we improve the accuracy of dialogue state tracking methods of spoken dialogue systems. Our approach limits the effect of automatic speech recognition (ASR) errors. We incrementally enhance our interpretable rule-based core by complex neural networks. The resulting system achieves several published state-of-the-art results on public datasets. 2) Effective dialogue management in open domain dialogue is a difficult prob- lem, which highlights the challenges of natural language processing. In this thesis, we propose a principal solution to develop dialogue systems in open domains. The key idea of our approach is building dialogue systems which interactively learn from dialogues with users. The interactive learning enables the system to improve and to extend its knowledge base continually. As a part of this...
|
|
|
Novel Methods for Natural Language Generation in Spoken Dialogue Systems
Dušek, Ondřej ; Jurčíček, Filip (advisor) ; Ircing, Pavel (referee) ; Žabokrtský, Zdeněk (referee)
Title: Novel Methods for Natural Language Generation in Spoken Dialogue Systems Author: Ondřej Dušek Department: Institute of Formal and Applied Linguistics Supervisor: Ing. Mgr. Filip Jurčíček, Ph.D., Institute of Formal and Applied Linguistics Abstract: This thesis explores novel approaches to natural language generation (NLG) in spoken dialogue systems (i.e., generating system responses to be presented the user), aiming at simplifying adaptivity of NLG in three respects: domain portability, language portability, and user-adaptive outputs. Our generators improve over state-of-the-art in all of them: First, our gen- erators, which are based on statistical methods (A* search with perceptron ranking and sequence-to-sequence recurrent neural network architectures), can be trained on data without fine-grained semantic alignments, thus simplifying the process of retraining the generator for a new domain in comparison to previous approaches. Second, we enhance the neural-network-based gener- ator so that it takes preceding dialogue context into account (i.e., user's way of speaking), thus producing user-adaptive outputs. Third, we evaluate sev- eral extensions to the neural-network-based generator designed for producing output in morphologically rich languages, showing improvements in Czech generation. In...
|
|
|
User simulation for statistical dialogue systems
Michlíková, Vendula ; Jurčíček, Filip (advisor) ; Žabokrtský, Zdeněk (referee)
The purpose of this thesis is to develop and evaluate user simulators for a spoken dialogue system. Created simulators are operating on dialogue act level. We implemented a bigram simulator as a baseline system. Based on the baseline simulator, we created another bigram simulator that is trained on dialogue acts without slot values. The third implemented simulator is similar to an implemen- tation of a dialogue manager. It tracks its dialogue state and learns a dialogue strategy based on the state using supervised learning. The user simulators are implemented in Python 2.7, in ALEX framework for dialogue system development. Simulators are developed for PTICS application which operates in the domain of public transport information. Simulators are trained and evaluated using real human-machine dialogues collected with PTICS application. 1
|
|
|
Neural networks for automatic speaker, language, and sex identification
Do, Ngoc ; Jurčíček, Filip (advisor) ; Peterek, Nino (referee)
Title: Neural networks for automatic speaker, language, and sex identifica- tion Author: Bich-Ngoc Do Department: Institute of Formal and Applied Linguistics Supervisor: Ing. Mgr. Filip Jurek, Ph.D., Institute of Formal and Applied Linguistics and Dr. Marco Wiering, Faculty of Mathematics and Natural Sciences, University of Groningen Abstract: Speaker recognition is a challenging task and has applications in many areas, such as access control or forensic science. On the other hand, in recent years, deep learning paradigm and its branch, deep neural networks have emerged as powerful machine learning techniques and achieved state-of- the-art in many fields of natural language processing and speech technology. Therefore, the aim of this work is to explore the capability of a deep neural network model, recurrent neural networks, in speaker recognition. Our pro- posed systems are evaluated on TIMIT corpus using speaker identification task. In comparison with other systems in the same test conditions, our systems could not surpass reference ones due to the sparsity of validation data. In general, our experiments show that the best system configuration is a combination of MFCCs with their dynamic features and a recurrent neural network model. We also experiment recurrent neural networks and convo- lutional neural...
|
|
|
Development of trainable policies for spoken dialogue systems
Le, Thanh Cong ; Jurčíček, Filip (advisor) ; Peterek, Nino (referee)
Abstract Development of trainable policies for spoken dialogue systems Thanh Le In humanhuman interaction, speech is the most natural and effective manner of communication. Spoken Dialogue Systems (SDS) have been trying to bring that high level interaction to computer systems, so with SDS, you could talk to machines rather than learn to use mouse and keyboard for performing a task. However, as inaccuracy in speech recognition and inherent ambiguity in spoken language, the dialogue state (user's desire) can never be known with certainty, and therefore, building such a SDS is not trivial. Statistical approaches have been proposed to deal with these uncertainties by maintaining a probability distribution over every possible dialogue state. Based on these distributions, the system learns how to interact with users, somehow to achieve the final goal in the most effective manner. In Reinforcement Learning (RL), the learning process is understood as optimizing a policy of choosing action conditioned on the current belief state. Since the space of dialogue...
|
|
|
Rozpoznávání řeči pomocí KALDI
Plátek, Ondřej ; Jurčíček, Filip (advisor) ; Peterek, Nino (referee)
The topic of this thesis is to implement efficient decoder for speech recognition training system ASR Kaldi (http://kaldi.sourceforge.net/). Kaldi is already deployed with decoders, but they are not convenient for dialogue systems. The main goal of this thesis to develop a real time decoder for a dialogue system, which minimize latency and optimize speed. Methods used for speeding up the decoder are not limited to multi-threading decoding or usage of GPU cards for general computations. Part of this work is devoted to training an acoustic model and also testing it in the "Vystadial" dialogue system. Powered by TCPDF (www.tcpdf.org)
|
|
|
Development of a cloud platform for automatic speech recognition
Klejch, Ondřej ; Jurčíček, Filip (advisor) ; Bojar, Ondřej (referee)
This thesis presents a cloud platform for automatic speech recognition, CloudASR, built on top of Kaldi speech recognition toolkit. The platform sup- ports both batch and online speech recognition mode and it has an annotation interface for transcription of the submitted recordings. The key features of the platform are scalability, customizability and easy deployment. Benchmarks of the platform show that the platform achieves comparable performance with Google Speech API in terms of latency and it can achieve better accuracy on limited domains. Furthermore, the benchmarks show that the platform is able to handle more than 1000 parallel requests given enough computational resources. 1
|
guest ::
