Hagai Aronowitz

PhD in Computer Science

Currently a Post-doc in IBM T.J. Watson Research Center

e-mail: aronowitzh@yahoo.com

Research Interests:

  • Speech-to-Text  (GALE project)
  • Speech recognition
  • Speaker recognition
  • Speaker segmentation and clustering
  • Speech & Audio indexing
  • Pattern matching
  • Text classification
  • Searching for proximity in high-dimensional spaces.

Courses I teach

Publications (2004-2007) – all are online (See an updated list in http://aronowitzh.googlepages.com)

2004

[1]      Aronowitz H., Burshtein D., Amir A., "Speaker indexing in audio archives using test utterance Gaussian mixture modeling", in Proc. ICSLP, pp. 609-612, 2004.

 

[2]      Aronowitz H., Burshtein D., Amir A., "Text independent speaker recognition using speaker dependent word spotting ", in Proc. ICSLP, pp. 1789-1792, 2004.

 

2005

 

[3]      Aronowitz H., Burshtein D., Amir A., "Speaker indexing in audio archives using Gaussian mixture scoring simulation ", in “Machine learning for multimodal interaction: first international workshop, MLMI'04, revised selected papers”, pp. 243-250, 2005.

 

[4]      Aronowitz H.,  Burshtein D., Amir A., "A session-GMM generative model using test utterance Gaussian mixture modeling for speaker verification",  in Proc. ICASSP 2005.

 

[5]      Aronowitz H. Irony D., “Modeling intra-speaker variability for improved speaker recognition”, in SLSF’, 2005.

 

[6]      Goldberger J. and Aronowitz H., "A distance measure between GMMs based on the  unscented transform and its application to speaker recognition" , in Proc. Interspeech 2005.

 

[7]      Aronowitz H., Irony D., Burshtein D., "Modeling Intra-Speaker Variability for Speaker Recognition",  in Proc. Interspeech 2005.

 

[8]      Aronowitz H., Burshtein D., "Efficient Speaker Identification and Retrieval", in Proc. Interspeech 2005.

 

2006

 

[9]      Noor E., Aronowitz H., "Efficient language Identification using Anchor Models and Support Vector Machines",  in Proc. Odyssey, 2006.

 

[10]      Y. Qin, Q. Shi, Y.Y. Liu, H. Aronowitz, S. M. Chu, H-K. Kuo, and G. Zweig, “Advances in Mandarin Broadcast Speech Transcription at IBM under the DARPA GALE Program”, in Proc. ISCSLP, 2006.

 

2007

 

[11]      Aronowitz H., “Segmental modeling for speech segmentation”, in Proc. ICASSP 2007.

 

[12]      Aronowitz H., Burshtein D., “Efficient Speaker Recognition Using Approximated Cross Entropy (ACE)”,  in IEEE Trans. on Audio, Speech & Language Processing, September 2007.

 

[13]      Aronowitz H., “Speaker Recognition using Kernel-PCA and Intersession Variability Modeling”, in Proc. Interspeech, 2007.

 

[14]      Aronowitz H., “Trainable Speaker Diarization”, in Proc. Interspeech, 2007.

 

Patents Issued

  1. "Speaker recognition using dynamic time warp template spotting",

       U.S. Patent No. 7,050,973, May 23, 2006.

Patents Filed

  1. "Decreasing noise sensitivity in speech processing under adverse conditions",

 U.S. Patent Application No. 20030033143, filed August 13, 2001.

  1. "Method and apparatus for adapting reference templates",

       U.S. Patent Application No. 20040122669, filed December 24, 2002.

  1. "Phoneme lattice construction and its application to speech recognition and keyword spotting",

       U.S. Patent Application No. 20050010412, filed July 7, 2003.

  1. "Apparatus and methods for pronunciation lexicon compression",

             filed February 2004.