8th European Conference on Speech Communication and Technology

Geneva, Switzerland
September 1-4, 2003


Confidence Measure Driven Scalable Two-Pass Recognition Strategy for Large List Grammars

Miroslav Novak (1), Diego Ruiz (2)

(1) IBM T.J. Watson Reseach Center, USA
(2) Universite Catholique de Louvain, Belgium

In this article we will discuss recognition performance on large list grammars, a class of tasks often encountered in telephony applications. In these tasks, the user makes a selection from a large list of choices (e.g. stock quotes, yellow pages, etc). Though the redundancy of the complete utterance is often high enough to achieve high recognition accuracy, large search space presents a challenge for the recognizer, in particular, when real-time, low latency performance is required. We propose a confidence measure driven two-pass search strategy, exploiting the high mutual information between grammar states to improve pruning efficiency while minimizing the need for memory.

Full Paper

Bibliographic reference.  Novak, Miroslav / Ruiz, Diego (2003): "Confidence measure driven scalable two-pass recognition strategy for large list grammars", In EUROSPEECH-2003, 1157-1160.