Short news

Shakti: Farewell in Brno, next stop: Cambridge

This Wednesday, January 23rd, the stay of Shakti P. Rath (Ph.D. IIT Madras) at BUT Speech@FIT came to end. Shakti was with the group 2 years, from January 2011, supported by the SoMoPro program co-funded by the EC (Marie Curie actions) and the South-Moravian region.

Shakti came with a sound knowledge of normalization and adaptation techniques for automatic speech recognition (ASR), especially the vocal tract length normalization. In Brno, he worked on four important ASR topics:

  • Regional-feature maximum likelihood linear transforms (R-FMLLR), allowing to compensate inter-speaker variabilities with the help of class-specific feature-space transformation.
  • Speaker adaptation of Deep Neural Networks DNN Adaptation – this work was done in cooperation with Daniel Povey from Johns Hopkins University (Maryland, USA) mainly during Shakti’s visit to JHU’s Center for Language and Speech Processing.
  • Factorized FMLLR (QR-FMLLR) that allows to modify the number of adaptation parameters depending on the size of available training data.
  • and he started works on Regional vocal tract length normalization (R-VTLN) where similar approaches as in R-FMLLR are applied to train multiple-parameter VTLN by considering acoustic-class specific frequency warping.

Shakti’s work significantly advances the field of acoustic modeling and adaptations for speech recognition systems. During his stay in Brno, Shakti published his results at the prestigious Interspeech conference in Portland (Oregon, USA) and several other publications are in the pipeline.

ASRU 2013 in Olomouc

The ASRU workshop meets every two years and has a tradition of bringing together researchers from academia and industry in an intimate and collegial setting to discuss problems of common interest in automatic speech recognition and understanding. ASRU 2013 is organized by Brno University of Technology BUT Speech@FIT group. It will take place in Olomouc.

ASRU website: www.asru2013.org

Ondrej receives Josef Hlavka Prize

Ondra Glembek was awarded a Josef Hlávka Prize by foundation "Nadání (talent) Josefa, Marie a Zdeňky Hlávkových". Ondra took part at the ceremony held on 16.11.2012 in the Chateau Lužany u Přeštic. „Nadání Josefa, Marie a Zdeňky Hlávkových“ is the oldest Czech foundation supporting science and arts. It was founded in 1904 by Josef Hlávka (1831 – 1908) – an important Czech architect, builder, and philanthropist. The Josef Hlávka Prize is awarded to the best students and graduates of Prague public universities, Brno University of Technology and young talented scientists of the Academy of Sciences of the Czech Republic.

Dr. Ondrej

Ondrej Glembek successfully defended his PhD thesis "Optimization of Gaussian Mixture Subspace Models and Related Scoring Algorithms in Speaker Verification" on November 13, 2012. Congratulations, Dr. Ondrej!

Oponents: Niko Brümmer, Ph.D. (Agnitio Research, South Africa), William M. Campbell, Ph.D. (MIT Lincoln Laboratory, USA)

Phonexia Android Speaker ID App awarded at DevFest 2012 Prague

Phonexia Speaker ID App running on the mobile operating system Android was awarded at DevFest held on November 10 2012 in Prague. DevFest is an event supported by Google and in its "GugCamp" part, it allows companies and individuals to present innovative solutions based on Google technologies. Phonexia Speaker ID App is based on the theory and algorithms investigated in the BUT Speech@FIT group, the cooperation is supported by "Technologies of speech processing for efficient human-machine communication" project funded by Technology Agency of the Czech Republic. The origins of Brno speaker recognition development for mobiles can be traced down to the EC-sponsored MOBIO project.

Speech@FIT at Gaudeamus 2012

Our research group was presenting "serious science" as a part of Brno University of Technology at Gaudeamus 2012. This annually held exhibition is the best known education and study abroad fair in Central Europe. This year more than 30000 visitors came to see the exhibition. Students were asked to scream every hour and the loudest screamer won a small price, such as a mug, cap or bag. The loudest voice reached approximately 122dB from the one-meter-long distance.

The second BOSARIS workshop (2012) started - 14 people arrived!

We proudly announce that the second season of BOSARIS - the Brno Speaker Recognition Workshop - has recently started. (BOSARIS 2010 information page is here.)

The workshop is held at the Faculty of Information Technology, Brno University of Technology, Czech Republic. It lasts from October 29 until November 26, 2012.

The following people arrived to jointly work on the topic of robust speaker recognition:

Dr. Tomas

On October 2nd, Tomas Mikolov successfully defended his PhD thesis entitled "Statistical Language Models Based on Neural Networks". His work was supervised by Jan Cernocky.

Congratulations, Tomas!

Lukas is back

Lukas Burget, the research director and 'brain' of the group is back from 16-month sabbatical stay in SRI International. Lukas was working with people from SRI's speech group: Nicolas Scheffer, Lu Ferrer, Yun Lei, and others, mainly on the BEST project supported by IARPA. See Lukas's web pages for common publications with SRI folks.

Kaldi workshop in Brno

We are happy that Kaldi workshop has started at BUT in Brno this week. There are 13 people participating and working hard for the next 11 days. Among others we invited guys such as:
Dan Povey and Sanjeev Khudanpur from JHU, USA
Arnab Ghoshal from SoI Edinburgh, UK
Navdeep Jaitly from UoT Toronto, Canada
Thang Vu from KIT, Germany
Petr Motlicek from IDIAP, Switzerland