MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Design of the CMU Sphinx-4 Decoder Lamere, P.; Kwok, P.; Walker, W.; Gouva, E.; Singh, R.; Raj, B.; Wolf, P. TR2003-110 August 2003 Abstract The decoder of the sphinx-4 speech recognition system incorporates several new design strate- gies which have not been used earlier in conventional decoders of HMM-based large vocabulary speech recognition systems. Some new design aspects include graph construction for multi- level parallel decoding with independent simultaneous feature streams without the use of com- pound HMMs, the incorporation of a generalized search algorithm that subsumes Viterbi and full-forward decoding as special cases, design of generalized language HMM graphs from gram- mars and language models of multiple standard formats, that toggles trivially from ﬂat search structure to tree search structure etc. This paper describes some salient design aspects of the Sphinx-4 decoder and includes preliminary performance measures relating to speed and accu- racy. Eurospeech 2003 This work may not be copied or reproduced in whole or in part for any commercial purpose. Permission to copy in whole or in part without payment of fee is granted for nonproﬁt educational and research purposes provided that all such whole or partial copies include the following: a notice that such copying is by permission of Mitsubishi Electric Research Laboratories, Inc.; an acknowledgment of the authors and individual contributions to the work; and all applicable portions of the copyright notice. Copying, reproduction, or republishing for any other purpose shall require a license with payment of fee to Mitsubishi Electric Research Laboratories, Inc. All rights reserved. Copyright c  Mitsubishi Electric Research Laboratories, Inc., 2003 201 Broadway, Cambridge, Massachusetts 02139