Technical Program

ASR-3: Speech Recognition 3 and Synthesis

Session Type: Poster
Poster Time: Friday, December 16, 10:30 - 12:30
Location: Lahaina Bay Room
Session Chair: Pedro Moreno, Google Inc.
 
ASR-3.1: CODE-SWITCHING DETECTION USING MULTILINGUAL DNNS
         Emre Yilmaz; Radboud University Nijmegen, Netherlands
         Henk Van den Heuvel; Radboud University Nijmegen, Netherlands
         David Van Leeuwen; Radboud University Nijmegen, Netherlands
 
ASR-3.2: ATTRIBUTE BASED SHARED HIDDEN LAYERS FOR CROSS-LANGUAGE KNOWLEDGE TRANSFER
         Vipul Arora; University of Oxford, United Kingdom
         Aditi Lahiri; University of Oxford, United Kingdom
         Henning Reetz; Goethe University, Germany
 
ASR-3.3: TOWARDS ACOUSTIC MODEL UNIFICATION ACROSS DIALECTS
         Mohamed Elfeky; Google Inc., United States
         Meysam Bastani; Google Inc., United States
         Xavier Velez; Google Inc., United States
         Pedro Moreno; Google Inc., United States
         Austin Waters; Google Inc., United States
 
ASR-3.4: BOOSTING PERFORMANCE ON LOW-RESOURCE LANGUAGES BY STANDARD CORPORA: AN ANALYSIS
         František Grézl; Brno University of Technology, Czech Republic
         Martin Karafiat; Brno University of Technology, Czech Republic
 
ASR-3.5: MULTILINGUAL BLSTM AND SPEAKER-SPECIFIC VECTOR ADAPTATION IN 2016 BUT BABEL SYSTEM
         Martin Karafiat; Speech@FIT VUT, Czech Republic
         Murali Karthick Baskar; Speech@FIT VUT, Czech Republic
         Pavel Matějka; Speech@FIT VUT, Czech Republic
         Karel Vesely; Speech@FIT VUT, Czech Republic
         František Grézl; Speech@FIT VUT, Czech Republic
         Jan Černocký; Speech@FIT VUT, Czech Republic
 
ASR-3.6: DNN ADAPTATION FOR RECOGNITION OF CHILDREN SPEECH THROUGH AUTOMATIC UTTERANCE SELECTION
         Marco Matassoni; Fondazione Bruno Kessler, Italy
         Daniele Falavigna; Fondazione Bruno Kessler, Italy
         Diego Giuliani; Fondazione Bruno Kessler, Italy
 
ASR-3.7: LOW-RANK BASES FOR FACTORIZED HIDDEN LAYER ADAPTATION OF DNN ACOUSTIC MODELS
         Lahiru Samarakoon; National University of Singapore, Singapore
         Khe Chai Sim; Google Inc., United States
 
ASR-3.8: DEEP NEURAL NETWORK BASED ACOUSTIC MODEL PARAMETER REDUCTION USING MANIFOLD REGULARIZED LOW RANK MATRIX FACTORIZATION
         Hoon Chung; Electronics and Telecommunications Research Institute, Republic of Korea
         Jeom Ja Kang; Electronics and Telecommunications Research Institute, Republic of Korea
         Ki Young Park; Electronics and Telecommunications Research Institute, Republic of Korea
         Sung Joo Lee; Electronics and Telecommunications Research Institute, Republic of Korea
         Jeon Gue Park; Electronics and Telecommunications Research Institute, Republic of Korea
 
ASR-3.9: AUTOMATED STRUCTURE DISCOVERY AND PARAMETER TUNING OF NEURAL NETWORK LANGUAGE MODEL BASED ON EVOLUTION STRATEGY
         Tomohiro Tanaka; Tokyo Institute of Technology, Japan
         Takafumi Moriya; NTT Corporation, Japan
         Takahiro Shinozaki; Tokyo Institute of Technology, Japan
         Shinji Watanabe; Mitsubishi Electric Research Laboratories, United States
         Takaaki Hori; Mitsubishi Electric Research Laboratories, United States
         Kevin Duh; Johns Hopkins University, United States
 
ASR-3.10: ENTROPY-BASED PRUNING OF HIDDEN UNITS TO REDUCE DNN PARAMETERS
         Gautam Mantena; National University of Singapore, Singapore
         Khe Chai Sim; Google Inc., United States
 
ASR-3.11: INFLUENCE OF CORPUS SIZE AND CONTENT ON THE PERCEPTUAL QUALITY OF A UNIT SELECTION MARYTTS VOICE
         Florian Hinterleitner; TU Berlin, Germany
         Benjamin Weiss; TU Berlin, Germany
         Sebastian Möller; TU Berlin, Germany
 
ASR-3.12: MEDIAN-BASED GENERATION OF SYNTHETIC SPEECH DURATIONS USING A NON-PARAMETRIC APPROACH
         Srikanth Ronanki; University of Edinburgh, United Kingdom
         Oliver Watts; University of Edinburgh, United Kingdom
         Simon King; University of Edinburgh, United Kingdom
         Gustav Henter; University of Edinburgh, United Kingdom
 
ASR-3.13: F0 TRANSFORMATION TECHNIQUES FOR STATISTICAL VOICE CONVERSION WITH DIRECT WAVEFORM MODIFICATION WITH SPECTRAL DIFFERENTIAL
         Kazuhiro Kobayashi; Nara Institute of Science and Technology, Japan
         Tomoki Toda; Nagoya University, Japan
         Satoshi Nakamura; Nara Institute of Science and Technology, Japan