Learn More
In the early 1990s, the availability of the TIMIT read-speech phonetically transcribed corpus led to work at AT&T on the automatic inference of pronunciation variation. This work, brie¯y summarized here, used stochastic decision trees trained on phonetic and linguistic features, and was applied to the DARPA North American Business News read-speech ASR task.(More)
We describe a lattice generation method that produces high-quality lattices with less than 10% increased computation over standard Viterbi decoding. Using the North American Business News (NAB) task, we show our method is within 0.2% in lattice word-error rate of 'full lattices', which are those that contain all the recognition hypotheses within the search(More)
We combine our earlier approach to context-dependent network representation with our algorithm for determinizing weighted networks to build optimized networks for large-vocabulary speech recognition combining an n-gram language model, a pronunciation dictionary and context-dependency modeling. While fully-expanded networks have been used before in(More)
We describe the AT&T recognition system used in the Large Vocabulary Conversational Speech Recognition (LVCSR-2000) evaluation. It is based on multi-pass rescoring of weighted Finite State Machines (FSMs) using progressively more accurate models. The final stages, which provide highest recognition accuracy , use combined outputs of several models that were(More)