Technical Program

M1: Automatic Speech Recognition I

Session Type: Poster
Poster Time: Monday, December 14, 11:00 - 12:30
Location: Earth + Air + Fire
Session Chair: Najim Dehak, Massachusetts Institute of Technology
 
M1.1: DIFFERENT WORD REPRESENTATIONS AND THEIR COMBINATION FOR PROPER NAME RETRIEVAL FROM DIACHRONIC DOCUMENTS
         Irina Illina; LORIA-INRIA, France
         Dominique Fohr; LORIA-INRIA, France
 
M1.2: SPARSE NON-NEGATIVE MATRIX LANGUAGE MODELING FOR GEO-ANNOTATED QUERY SESSION DATA
         Ciprian Chelba; Google Inc, United States
         Noam Shazeer; Google Inc, United States
 
M1.3: TRAINING DATA PSEUDO-SHUFFLING AND DIRECT DECODING FRAMEWORK FOR RECURRENT NEURAL NETWORK BASED ACOUSTIC MODELING
         Naoyuki Kanda; National Institute of Information and Communications Technology, Japan
         Mitsuyoshi Tachimori; National Institute of Information and Communications Technology, Japan
         Xugang Lu; National Institute of Information and Communications Technology, Japan
         Hisashi Kawai; National Institute of Information and Communications Technology, Japan
 
M1.4: ON CONSTRUCTING AND ANALYSING AN INTERPRETABLE BRAIN MODEL FOR THE DNN BASED ON HIDDEN ACTIVITY PATTERNS
         Khe Chai Sim; National University of Singapore, Singapore
 
M1.5: SPEAKER LOCATION AND MICROPHONE SPACING INVARIANT ACOUSTIC MODELING FROM RAW MULTICHANNEL WAVEFORMS
         Tara Sainath; Google Inc, United States
         Ron Weiss; Google Inc, United States
         Kevin Wilson; Google Inc, United States
         Arun Narayanan; Google Inc, United States
         Michiel Bacchiani; Google Inc, United States
         Andrew Senior; Google Inc, United States
 
M1.6: HYBRID DNN-LATENT STRUCTURED SVM ACOUSTIC MODELS FOR CONTINUOUS SPEECH RECOGNITION
         Suman Ravuri; International Computer Science Institute; University of California - Berkeley, United States
 
M1.7: DISCRIMINATIVE TRAINING OF CONTEXT-DEPENDENT LANGUAGE MODEL SCALING FACTORS AND INTERPOLATION WEIGHTS
         Shuangyu Chang; Microsoft Corporation, United States
         Abhik Lahiri; Microsoft Corporation, United States
         Issac Alphonso; Microsoft Corporation, United States
         Barlas Oguz; Microsoft Corporation, United States
         Michael Levit; Microsoft Corporation, United States
         Benoit Dumoulin; Facebook Inc., United States
 
M1.8: ACOUSTIC MODEL TRAINING BASED ON NODE-WISE WEIGHT BOUNDARY MODEL INCREASING SPEED OF DISCRETE NEURAL NETWORKS
         Ryu Takeda; Osaka University, Japan
         Kazunori Komatani; Osaka University, Japan
         Kazuhiro Nakadai; Honda Research Institute Japan Co., Ltd., Japan
 
M1.9: TWO-STAGE ASGD FRAMEWORK FOR PARALLEL TRAINING OF DNN ACOUSTIC MODELS USING ETHERNET
         Zhichao Wang; Institute of Acoustics, Chinese Academy of Sciences, China
         Xingyu Na; Institute of Acoustics, Chinese Academy of Sciences, China
         Xin Li; Institute of Acoustics, Chinese Academy of Sciences, China
         Jielin Pan; Institute of Acoustics, Chinese Academy of Sciences, China
         Yonghong Yan; Institute of Acoustics, Chinese Academy of Sciences, China
 
M1.10: RNNDROP: A NOVEL DROPOUT FOR RNNS IN ASR
         Taesup Moon; Samsung Advanced Institute of Technology, Republic of Korea
         Heeyoul Choi; Samsung Advanced Institute of Technology, Republic of Korea
         Hoshik Lee; Samsung Advanced Institute of Technology, Republic of Korea
         Inchul Song; Samsung Advanced Institute of Technology, Republic of Korea
 
M1.11: SPECTRAL LEARNING WITH NON NEGATIVE PROBABILITIES FOR FINITE STATE AUTOMATON
         Hadrien Glaude; Thales Airborne Systems / University Lille 1, France
         Cyrille Enderli; Thales Airborne Systems, France
         Olivier Pietquin; University Lille 1, France
 
M1.12: DEEP BI-DIRECTIONAL RECURRENT NETWORKS OVER SPECTRAL WINDOWS
         Abdel-Rahman Mohamed; Microsoft, United States
         Frank Seide; Microsoft, United States
         Dong Yu; Microsoft, United States
         Jasha Droppo; Microsoft, United States
         Andreas Stolcke; Microsoft, United States
         Geoffrey Zweig; Microsoft, United States
         Gerald Penn; University of Toronto, Canada
 
M1.13: PERSONALIZING UNIVERSAL RECURRENT NEURAL NETWORK LANGUAGE MODEL WITH USER CHARACTERISTIC FEATURES BY SOCIAL NETWORK CROWDSOURCING
         Bo-Hsiang Tseng; National Taiwan University, Taiwan
         Hung-yi Lee; National Taiwan University, Taiwan
         Lin-Shan Lee; National Taiwan University, Taiwan
 
M1.14: TIME DELAY DEEP NEURAL NETWORK-BASED UNIVERSAL BACKGROUND MODELS FOR SPEAKER RECOGNITION
         David Snyder; The Johns Hopkins University, United States
         Daniel Garcia-Romero; The Johns Hopkins University, United States
         Daniel Povey; The Johns Hopkins University, United States