Combined Optimisation of Baseforms and Model Parameters in Speech Recognition Based on Acoustic Subword Units
نویسنده
چکیده
A major challenge in speech recognition is creating a lexicon which is robust to inter-and intra-speaker variations. This is even more so in speech recognisers based on non-linguistic units, e.g., acoustic subword units (ASWUs), since no standard pronunciation dictionaries are available. Thus the baseforms describing the vocabulary words in terms of the recognition units need to be generated from training data. In this paper we propose an algorithm for ASWU-based speech recognition which performs a combined optimisation of the baseforms and the subword models. The resulting system has been tested on the DARPA Resource Management task, and is shown to perform comparable to a baseline phoneme based system.
منابع مشابه
Incorporating linguistic knowledge and automatic baseform generation in acoustic subword unit based speech recognition
A major challenge in speech recognition based on acoustic subword units is creating a lexicon which is robust to interand intra-speaker variations. In this paper we present two di erent approaches for incorporating simple word-level linguistic knowledge into the labelling step of the training procedure. The proposed systems also utilise a scheme for combined optimisation of baseforms and subwor...
متن کاملCombined Optimisation of Baseforms and Subword Models for an Hmm Based Speech Recogniser
In this paper a framework for combined optimisation of baseforms and subword models for a speech recogniser is proposed. Given a set of subword Hidden Markov Models (HMMs) and a set of utterances of a speciic word, the modiied tree-trellis algorithmand the Baum-Welch re-estimation procedure is used iteratively to achieve a combined optimisation of baseforms and sub-word models. The DARPA Resour...
متن کاملSpeech recognition using automatically derived acoustic baseforms
This paper investigates procedures for obtaining user-con gurable speech recognition vocabularies. These procedures use example utterances of vocabulary words to perform unsupervised automatic acoustic baseform determination in terms of a set of speaker independent subword acoustic units. Several procedures, di ering both in the de nition of subword acoustic model context and in the phonotactic...
متن کاملTraining of Lexica for Subword-Based Speech Recognisers
In this paper we present an automatic optimal baseform determination algorithm. Given a set of subword Hidden Markov Models (HMMs) and acoustic tokens of a speciic word, we apply the tree-trellis N-best search algorithm to nd the optimal baseforms (transcriptions) in the maximum likelihood sense. The proposed algorithm is used in an iterative manner, creating a series of lexica trained from the...
متن کاملA Joint Segmentation and Labelling Scheme for use inAcoustic
A major challenge in speech recognition based on acoustic subword units is creating a lexicon which is robust to inter-and intra-speaker variations. In this paper we present a joint seg-mentation and labelling scheme to incorporate word-level linguistic knowledge into the training procedure. The proposed system is also based on a combined optimisation of the base-forms and the subword models. F...
متن کامل