• DocumentCode
    672353
  • Title

    The IBM keyword search system for the DARPA RATS program

  • Author

    Mangu, Lidia ; Soltau, Hagen ; Hong-Kwang Kuo ; Saon, George

  • Author_Institution
    IBM T. J. Watson Res. Center, Yorktown Heights, NY, USA
  • fYear
    2013
  • fDate
    8-12 Dec. 2013
  • Firstpage
    204
  • Lastpage
    209
  • Abstract
    The paper describes a state-of-the-art keyword search (KWS) system in which significant improvements are obtained by using Convolutional Neural Network acoustic models, a two-step speech segmentation approach and a simplified ASR architecture optimized for KWS. The system described in this paper had the best performance in the 2013 DARPA RATS evaluation for both Levantine and Farsi.
  • Keywords
    neural nets; query processing; speech recognition; ASR architecture; DARPA RATS program; IBM keyword search system; KWS system; automatic speech recognition; convolutional neural network acoustic model; speech segmentation; Acoustics; Decoding; Hidden Markov models; Keyword search; Lattices; Rats; Speech; audio indexing; keyword spotting; spoken term detection; system combination;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Automatic Speech Recognition and Understanding (ASRU), 2013 IEEE Workshop on
  • Conference_Location
    Olomouc
  • Type

    conf

  • DOI
    10.1109/ASRU.2013.6707730
  • Filename
    6707730