- I have been with Charles University in Prague since September 2011. Currently, I am an assistant professor at the In... moreI have been with Charles University in Prague since September 2011. Currently, I am an assistant professor at the Institute of Formal and Applied Linguistics. I am interested in development of statistical methods for dialogue systems to resolve the inherent uncertainty in spoken speech. My main research interests lie in applying reinforcement learning and Bayesian techniques to dialogue management and natural language processing.
Before joining Charles University, I worked as a research associate at Cambridge University. My work involved development of statistical methods for spoken dialogue management, in particular POMDPs (Partially-Observable Markov Decision Processes). Prior that I worked as a software consultant in IT industry. I participated in projects implementing internet/IP telephony and automation of customer services.
I received my PhD degree in 2007 from the department of Computer Science at the University of West Bohemia in Pilsen (Czech Republic) on the subject of semantic parsing for spoken dialogue systems. During my PhD studies, I spent 10 months as a visiting scholar in the Center for Language and Speech Processing at The Johns Hopkins University (USA). Prior that I obtained a master's degree in Discrete Mathematics focused on closure techniques in Hamiltonian graph theory and a master’s degree in Computer Science on the subject of large vocabulary speech decoders for Czech language.edit
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Abstrakt This paper describes a dialog system that enables university students to registrate for examinations using telephone. Presented dialog system serves as supplement to university web-based information system and enables all... more
Abstrakt This paper describes a dialog system that enables university students to registrate for examinations using telephone. Presented dialog system serves as supplement to university web-based information system and enables all student, including disabled ones, to easily make or cancel registration for their exams. System is based on VoiceXML and unreservedly use speech synthesis and recognition for system-driven telephone dialog. Whole system is designed as kind of framework that would simplify developing and testing ...
Research Interests:
Research Interests:
Research Interests:
Abstract This paper describes an implementation of a statistical semantic parser for a closed domain with limited amount of training data. We implemented the hidden vector state model, which we present as a structure discrimination of a... more
Abstract This paper describes an implementation of a statistical semantic parser for a closed domain with limited amount of training data. We implemented the hidden vector state model, which we present as a structure discrimination of a flat-concept model. The model was implemented in the graphical modeling toolkit. We introduced into the hidden vector state model a concept insertion penalty as a part of pattern recognition approach.
Abstract Reinforcement learning methods have been successfully used to optimise dialogue strategies in statistical dialogue systems. Typically, reinforcement techniques learn on-policy ie, the dialogue strategy is updated online while the... more
Abstract Reinforcement learning methods have been successfully used to optimise dialogue strategies in statistical dialogue systems. Typically, reinforcement techniques learn on-policy ie, the dialogue strategy is updated online while the system is interacting with a user. An alternative to this approach is off-policy reinforcement learning, which estimates an optimal dialogue strategy offline from a fixed corpus of previously collected dialogues.
Spoken dialogue systems need to be able to interpret the spoken input from the user. This is done by mapping the user's spoken utterance to a representation of the meaning of that utterance, and then passing this representation to the... more
Spoken dialogue systems need to be able to interpret the spoken input from the user. This is done by mapping the user's spoken utterance to a representation of the meaning of that utterance, and then passing this representation to the dialogue manager. This process begins with the application of automatic speech recognition (ASR) technology, which maps the speech to hypotheses about the sequence of words in the utterance.
Abstract This paper presents a semantic parser that transforms an initial semantic hypothesis into the correct semantics by applying an ordered list of transformation rules. These rules are learnt automatically from a training corpus with... more
Abstract This paper presents a semantic parser that transforms an initial semantic hypothesis into the correct semantics by applying an ordered list of transformation rules. These rules are learnt automatically from a training corpus with no prior linguistic knowledge and no alignment between words and semantic concepts. The learning algorithm produces a compact set of rules which enables the parser to be very efficient while retaining high accuracy.
WP/Task responsible UCAM Other contributors HWU, UEDIN Author (s) Filip Jurcıcek, Simon Keizer, François Mairesse, Kai Yu, Steve Young, Srinivanan Janarthanam, Helen Hastie, Xingkun Liu, and Oliver Lemon EC Project Officer Philippe Gelin... more
WP/Task responsible UCAM Other contributors HWU, UEDIN Author (s) Filip Jurcıcek, Simon Keizer, François Mairesse, Kai Yu, Steve Young, Srinivanan Janarthanam, Helen Hastie, Xingkun Liu, and Oliver Lemon EC Project Officer Philippe Gelin Keywords Spoken dialogue systems, dialogue management, user simulation
Page 1. D1.4.2: Summary-based policy optimisation F. Jurcıcek, M. Gašic, B. Thomson, F. Lef`evre, S. Young Distribution: Public CLASSiC Computational Learning in Adaptive Systems for Spoken Conversation 216594 Deliverable 1.4.2 February... more
Page 1. D1.4.2: Summary-based policy optimisation F. Jurcıcek, M. Gašic, B. Thomson, F. Lef`evre, S. Young Distribution: Public CLASSiC Computational Learning in Adaptive Systems for Spoken Conversation 216594 Deliverable 1.4.2 February 2010 Project funded by the European Community under the Seventh Framework Programme for Research and Technological Development The deliverable identification sheet is to be found on the reverse of this page. Page 2. Project ref.
Abstract The LVCSR systems are developed for several decades. Nowadays LVCSR system take advantage of increasing performance of computers and sophisticated algorithms. There is permanent effort to integrate larger lexicons into LVCSR... more
Abstract The LVCSR systems are developed for several decades. Nowadays LVCSR system take advantage of increasing performance of computers and sophisticated algorithms. There is permanent effort to integrate larger lexicons into LVCSR system. A n-gram language model can improve recognition accuracy. However, memory requirments are substantial, so in practise for real-time implementation of LVCSR system only a bigram language model is often considered.
Abstract:-In this paper a LVCSR system with implementation of the Czech voice assimilation phenomenon is proposed. The recognition system uses lexical trees and a bigram language model. The first part of this article is focused on voice... more
Abstract:-In this paper a LVCSR system with implementation of the Czech voice assimilation phenomenon is proposed. The recognition system uses lexical trees and a bigram language model. The first part of this article is focused on voice assimilation phenomenon description, triphone lexical tree construction, and voice assimilation impact on LVCSR system performance. The second part outlines lexical tree decoding algorithm based on Viterbi search with pruning.
Abstract. The key component of a spoken dialogue system is a spoken understanding module. There are many approaches to the understanding module design and one of the most perspective is a statistical based semantic parsing. This paper... more
Abstract. The key component of a spoken dialogue system is a spoken understanding module. There are many approaches to the understanding module design and one of the most perspective is a statistical based semantic parsing. This paper presents a combination of a set of modifications of the hidden vector state (HVS) parser which is a very popular method for the statistical semantic parsing.
Abstract While data-driven methods for spoken language understanding reduce maintenance and portability costs compared with handcrafted parsers, the collection of word-level semantic annotations for training remains a time-consuming task.... more
Abstract While data-driven methods for spoken language understanding reduce maintenance and portability costs compared with handcrafted parsers, the collection of word-level semantic annotations for training remains a time-consuming task. A recent line of research has focused on building generative models from unaligned semantic representations, using expectation-maximisation techniques to align semantic concepts.
This paper describes use of negative examples in training the HVS semantic model. We present a novel initialization of the lexical model using negative examples extracted automatically from a semantic corpus as well as description of an... more
This paper describes use of negative examples in training the HVS semantic model. We present a novel initialization of the lexical model using negative examples extracted automatically from a semantic corpus as well as description of an algorithm for extraction these examples. We evaluated the use of negative examples on a closed domain human-human train timetable dialogue corpus. We significantly improved the standard PARSEVAL scores of the baseline system.
Page 1. D1.5: Online adaptation of dialogue systems F. Jurcıcek, M. Gašic, S. Young, R. Laroche, G. Putois, M. Geist and O. Pietquin Distribution: Public CLASSiC Computational Learning in Adaptive Systems for Spoken Conversation 216594... more
Page 1. D1.5: Online adaptation of dialogue systems F. Jurcıcek, M. Gašic, S. Young, R. Laroche, G. Putois, M. Geist and O. Pietquin Distribution: Public CLASSiC Computational Learning in Adaptive Systems for Spoken Conversation 216594 Deliverable 1.5 February 2011 Project funded by the European Community under the Seventh Framework Programme for Research and Technological Development The deliverable identification sheet is to be found on the reverse of this page. Page 2. Project ref.
Spoken dialogue systems (SDS) allow users to interact with a wide variety of information systems using speech as the primary, and often the only, communication medium. The principal elements of an SDS are a speech understanding component... more
Spoken dialogue systems (SDS) allow users to interact with a wide variety of information systems using speech as the primary, and often the only, communication medium. The principal elements of an SDS are a speech understanding component which converts each spoken input into an abstract semantic representation called a user dialogue act (see Chap.
Abstract This paper discusses a usage of a mumble model in a Czech telephone dialogue system designed and constructed at the Department of Cybernetics, University of West Bohemia, and describes benefits of the mumble model to speech... more
Abstract This paper discusses a usage of a mumble model in a Czech telephone dialogue system designed and constructed at the Department of Cybernetics, University of West Bohemia, and describes benefits of the mumble model to speech recognition, namely to a rejection method. Firstly, the overview of the Czech telephone dialogue system and its recognition engine is given. The recognition is based on a statistical approach.
Abstract The partially observable Markov decision process (POMDP) provides a popular framework for modelling spoken dialogue. This paper describes how the expectation propagation algorithm (EP) can be used to learn the parameters of the... more
Abstract The partially observable Markov decision process (POMDP) provides a popular framework for modelling spoken dialogue. This paper describes how the expectation propagation algorithm (EP) can be used to learn the parameters of the POMDP user model. Various special probability factors applicable to this task are presented, which allow the parameters be to learned when the structure of the dialogue is complex.
Abstract This paper describes how Bayesian updates of dialogue state can be used to build a bus information spoken dialogue system. The resulting system was deployed as part of the 2010 Spoken Dialogue Challenge. The purpose of this paper... more
Abstract This paper describes how Bayesian updates of dialogue state can be used to build a bus information spoken dialogue system. The resulting system was deployed as part of the 2010 Spoken Dialogue Challenge. The purpose of this paper is to describe the system, and provide both simulated and human evaluations of its performance. In control tests by human users, the success rate of the system was 24.5% higher than the baseline Lets Go! system.
Reinforcement techniques have been successfully used to maximise the expected cumulative reward of statistical dialogue systems. Typically, reinforcement learning is used to estimate the parameters of a dialogue policy which selects the... more
Reinforcement techniques have been successfully used to maximise the expected cumulative reward of statistical dialogue systems. Typically, reinforcement learning is used to estimate the parameters of a dialogue policy which selects the system's responses based on the inferred dialogue state. However, the inference of the dialogue state itself depends on a dialogue model which describes the expected behaviour of a user when interacting with the system.
Abstract Modelling dialogue as a Partially Observable Markov Decision Process (POMDP) enables a dialogue policy robust to speech understanding errors to be learnt. However, a major challenge in POMDP policy learning is to maintain... more
Abstract Modelling dialogue as a Partially Observable Markov Decision Process (POMDP) enables a dialogue policy robust to speech understanding errors to be learnt. However, a major challenge in POMDP policy learning is to maintain tractability, so the use of approximation is inevitable. We propose applying Gaussian Processes in Reinforcement learning of optimal POMDP dialogue policies, in order (1) to make the learning process faster and (2) to obtain an estimate of the uncertainty of the approximation.
Abstract In real-world applications, modelling dialogue as a POMDP requires the use of a summary space for the dialogue state representation to ensure tractability. Sub-optimal estimation of the value function governing the selection of... more
Abstract In real-world applications, modelling dialogue as a POMDP requires the use of a summary space for the dialogue state representation to ensure tractability. Sub-optimal estimation of the value function governing the selection of system responses can then be obtained using a grid-based approach on the belief space. In this work, the Monte-Carlo control technique is extended so as to reduce training over-fitting and to improve robustness to semantic noise in the user input.
Abstract Statistical dialogue models have required a large number of dialogues to optimise the dialogue policy, relying on the use of a simulated user. This results in a mismatch between training and live conditions, and significant... more
Abstract Statistical dialogue models have required a large number of dialogues to optimise the dialogue policy, relying on the use of a simulated user. This results in a mismatch between training and live conditions, and significant development costs for the simulator thereby mitigating many of the claimed benefits of such models. Recent work on Gaussian process reinforcement learning, has shown that learning can be substantially accelerated.