Abstract
Heterophones pose challenges during training of automatic speech recognition (ASR) systems because they involve ambiguity in the pronunciation of an orthographic representation of a word. Heterophones are words that have the same spelling but different pronunciations. This paper addresses the problem of heterophonic languages by developing the concept of a Composite Phoneme (CP) as a basic pronunciation unit for speech recognition. A CP is a set of alternative sequences of phonemes. CP’s are developed specifically in the context of Arabic by defining phonetic units that are consonant centric and absorb phonemically contrastive short vowels and gemination, not represented in the Arabic Modern Orthography (MO). CPs alleviate the need to diacritize MO into Classical Orthography (CO), to represent short vowels and stress, before generating pronunciation in terms of Simple Phonemes (SP). We develop algorithms to generate CP pronunciation from MO, and SP pronunciation from CO to map a word into a single pronunciation. We investigate the performance of CP, SP, UG (Undiacritized Grapheme), and DG (Diacritized Grapheme) ASRs. The experimental results suggest that UG and DG are inferior to SP and CP. For the A-SpeechDB corpus with MO vocabulary of 8000, the WER for bigram and context dependent phone are: 11.78, 12.64, and 13.59 % for CP, SP_M (SP from manual diacritized CO), and SP_A (SP from automated diacritized MO) respectively. For vocabulary of 24,000 MO words, the corresponding WER’s are 13.69, 15.08, and 16.86 %. For uniform statistical model, SP has a lower WER than CP. For context independent phone (CI), CP has lower WER than SP.
Highlights
A standard automatic speech recognition (ASR) system consists of a language model (LM) that governs the sequence of words in an utterance, a dictionary that maps words into sequences of pronunciation units, and Hidden Markov Models (HMMs) corresponding to pronunciation units that stochastically model acoustic events (Huang and Acero 2001)
For languages with a deep orthography, the Grapheme approach can be inferior to the Simple Phonemes (SP) (Simple Phoneme) approach by up to 10 % word error rate (WER), depending on task and complexity of mapping from orthography to pronunciation (Kanthak and Ney 2002; Magimai-Doss et al 2003a, b)
To address the problem of HMM training for heterophonic languages, this paper develops the concept of a Composite Phoneme (CP)—a set of alternative sequences of phonemes, such as a syllable with multiple vowels choices
Summary
A standard automatic speech recognition (ASR) system consists of a language model (LM) that governs the sequence of words in an utterance, a dictionary that maps words into sequences of pronunciation units, and Hidden Markov Models (HMMs) corresponding to pronunciation units that stochastically model acoustic events (Huang and Acero 2001). The pronunciation units could correspond to either a phoneme (an individual phonetic segment) or to a syllable (a structured sequence of phonemes), with the choice depending on characteristics of a given language’s phonological system. In an ASR system that uses the phoneme, phonetic segments could correspond to O(N) context independent Monophones for a language with. Training of HMMs is conducted on phonetic transcriptions of speech utterances, which are derived from orthographic transcriptions using ortho-phonetic mapping. One important problem that arises during the training phase is the ambiguity posed by heterophones—words that have the same orthographic representation but different pronunciations (e.g., in English, the noun “bow”, referring to a weapon, and the verb “bow”, referring to a gesture of respect) (Wikipedia 2016)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
