Lattice-Free Sequence Discriminative Training for Phoneme-Based Neural Transducers
IEEE International Conference on Acoustics Speech and Signal Processing, pp. 1–5
Abstract
Recently, RNN-Transducers have achieved remarkable results on various automatic speech recognition tasks. However, lattice-free sequence discriminative training methods, which obtain superior performance in hybrid models, are rarely investigated in RNN-Transducers. In this work, we propose three lattice-free training objectives, namely lattice-free maximum mutual information, lattice-free segment-level minimum Bayes risk, and lattice-free minimum Bayes risk, which are used for the final posterior output of the phoneme-based neural transducer with a limited context dependency. Compared to criteria using N-best lists, lattice-free methods eliminate the decoding step for hypotheses generation during training, which leads to more efficient training. Experimental results show that lattice-free methods gain up to 6.5% relative improvement in word error rate compared to a sequence-level cross-entropy trained model. Compared to the N-best-list based minimum Bayes risk objectives, lattice-free methods gain 40% - 70% relative training time speedup with a small degradation in performance.
Authors 4
-
Affiliation as printed
RWTH Aachen University,Human Language Technology and Pattern Recognition,Computer Science Department,Aachen,Germany,52074
-
Affiliation as printed
RWTH Aachen University,Human Language Technology and Pattern Recognition,Computer Science Department,Aachen,Germany,52074
AppTek GmbH, Aachen, Germany
-
Affiliation as printed
RWTH Aachen University,Human Language Technology and Pattern Recognition,Computer Science Department,Aachen,Germany,52074
AppTek GmbH, Aachen, Germany
-
Affiliation as printed
RWTH Aachen University,Human Language Technology and Pattern Recognition,Computer Science Department,Aachen,Germany,52074
AppTek GmbH, Aachen, Germany
Cited by 3 stored of 3
3 results
No patents citing this paper on Lens.org (checked 2026-10-06).