Hi! I'm a PhD student interested in spoken dialog systems and machine learning.

I am interested in natural languages and human communication, particularly I am looking how to teach computers to understand the languages we speak and to talk back. Currently, I work in the area of spoken dialog systems and I am trying to look for ways to make them more reactive and pleasant for the users while keeping their design and development simple. To achieve that I am using the recurrent neural networks models, because they seem to be a very natural model that has been shown capable of learning complex relationships, such as those present in natural human communication. I am also interested in statistical learning theory and understanding the underlying mechanisms of why neural networks work.

Google Scholar Publications.

Research Projects

  LecTrack: LSTM Dialog State Tracker

A dialog state tracker is an important component in modern spoken dialog systems. We present the first trainable incremental dialog state tracker that directly uses automatic speech recognition hypotheses to track the state. It is based on a long short-term memory recurrent neural network, and it is fully trainable from annotated data. The tracker achieves promissing performance on the Method and Requested tracking sub-tasks in DSTC2.

(paper, old paper)

  ALEX spoken dialog system

ALEX is an open-source spoken dialog system/framework.

It powers our Public Transportation Information spoken dialog system, which searches metropolitian and country-wide public transport schedules and provides this information over voice. It runs at the telephone number 800 899 998 (you need to be in the Czech Republic to reach it).

  • We collected some Czech and English spoken data using our running systems and give it out Under the CC-BY-SA 3.0 License.
  • ALEX uses a simple but effective Bayesian belief state update described in our paper.
  • We described how we bootstrapped ALEX in a new domain in this paper.

  Explicit Semantic Analysis

We applied Explicit Semantic Analysis for automatic and cross-lingual link discovery in document collections.

Won 1st-3rd place at CLLD NTCIR-9 competition

(paper, thesis, code).


I am a PhD student at Institute of Formal and Applied Linguistics at Faculty of Mathematics and Physics at Charles University in Prague, Czech Republic. My research topic is modelling of complex statistical dialogue systems, and my thesis advisor is Filip Jurčíček.



  • 04/15/2015 - Dialog State Tracking with LSTM's @ IBM Watson Research, Prague
  • 02/23/2015 - LSTM dialog state tracker @ UFAL, MFF, Prague
  • 11/09/2014 - ALEX dialogue framework at Machine Learning Meet-ups in Prague (video, slides)