ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020
This paper describes new reference benchmark results based on hybrid Hidden Markov Model and Deep... more This paper describes new reference benchmark results based on hybrid Hidden Markov Model and Deep Neural Networks (HMM-DNN) for the GlobalPhone (GP) multilingual text and speech database. GP is a multilingual database of high-quality read speech with corresponding transcriptions and pronunciation dictionaries in more than 20 languages. Moreover, we provide new results for five additional languages, namely, Amharic, Oromo, Tigrigna, Wolaytta, and Uyghur. Across the 22 languages considered, the hybrid HMM-DNN models outperform the HMM-GMM based models regardless of the size of the training speech used. Overall, we achieved relative improvements that range from 7.14% to 59.43%.
ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020
In this work, we present speech recognition systems for four Ethiopian languages: Amharic, Tigrig... more In this work, we present speech recognition systems for four Ethiopian languages: Amharic, Tigrigna, Oromo and Wolaytta. We have used comparable training corpora of about 20 to 29 hours speech and evaluation speech of about 1 hour for each of the languages. For Amharic and Tigrigna, lexical and language models of different vocabulary size have been developed. For Oromo and Wolaytta, the training lexicons have been used for decoding. We achieved relative word error rate (WER) reductions for all the languages by using Deep Neural Networks (DNN) based acoustic models that range from 15.1% to 31.45%. The relative improvement obtained for Wolaytta speech recognition system is much higher (31.45%) than the improvement achieved for the other languages. This attributes to the weaker language model and the bigger size of training speech we used for Wolaytta.
In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian langua... more In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian languages (Amharic, Tigrigna, Oromo and Wolaytta). The aim of the analysis is to select speech data from GP for the development of multilingual Automatic Speech Recognition (ASR) system for the Ethiopian languages. To this end, phonetic overlaps among GP and Ethiopian languages have been analyzed. The result of our analysis shows that there is much phonetic overlap among Ethiopian languages although they are from three different language families. From GP, Turkish, Uyghur and Croatian are found to have much overlap with the Ethiopian languages. On the other hand, Korean has less phonetic overlap with the rest of the languages. Moreover, morphological complexity of the GP and Ethiopian languages, reflected by type to token ration (TTR) and out of vocabulary (OOV) rate, has been analyzed. Both metrics indicated the morphological complexity of the languages. Korean and Amharic have been identifie...
In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian langua... more In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian languages (Amharic, Tigrigna, Oromo and Wolaytta). The aim of the analysis is to select speech data from GP for the development of multilingual Automatic Speech Recognition (ASR) system for the Ethiopian languages. To this end, phonetic overlaps among GP and Ethiopian languages have been analyzed. The result of our analysis shows that there is much phonetic overlap among Ethiopian languages although they are from three different language families. From GP, Turkish, Uyghur and Croatian are found to have much overlap with the Ethiopian languages. Moreover, morphological complexity of the GP and Ethiopian languages, reflected by type to token ration (TTR) and out of vocabulary (OOV) rate, has been analyzed. Korean and Amharic have been identified as extremely morphologically complex compared to the other languages. Tigrigna, Russian, Turkish and Polish are also among the morphologically complex l...
In this paper, we describe the development of parallel corpora for Ethiopian Languages: Amharic, ... more In this paper, we describe the development of parallel corpora for Ethiopian Languages: Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. To check the usability of all the corpora we conducted baseline bi-directional statistical machine translation (SMT) experiments for seven language pairs. The performance of the bi-directional SMT systems shows that all the corpora can be used for further investigations. We have also shown that the morphological complexity of the Ethio-Semitic languages has a negative impact on the performance of the SMT especially when they are target languages. Based on the results we obtained, we are currently working towards handling the morphological complexities to improve the performance of statistical machine translation among the Ethiopian languages. * This work is licenced under a Creative Commons Attribution 4.0 International License. Page numbers and proceedings footer are added by the organizers. License details: http://creativecommons.org/licenses/b...
It is known that Automatic Speech Recognition (ASR) is very useful for human-computer interaction... more It is known that Automatic Speech Recognition (ASR) is very useful for human-computer interaction in all the human languages. However, due to its requirement for a big speech corpus, which is very expensive, it has not been developed for most of the languages. Multilingual ASR (MLASR) has been suggested to share existing speech corpora among related languages to develop an ASR for languages which do not have the required speech corpora. Literature shows that phonetic relatedness goes across language families. We have, therefore, conducted experiments on MLASR taking two language families: one as source (Oromo from Cushitic) and the other as target (Wolaytta from Omotic). Using Oromo Deep Neural Network (DNN) based acoustic model, Wolaytta pronunciation dictionary and language model we have achieved Word Error Rate (WER) of 48.34% for Wolaytta. Moreover, our experiments show that adding only 30 minutes of speech data from the target language (Wolaytta) to the whole training data (22....
ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021
The End-to-End (E2E) approach, which maps a sequence of input features into a sequence of graphem... more The End-to-End (E2E) approach, which maps a sequence of input features into a sequence of graphemes or words, to Automatic Speech Recognition (ASR) is a hot research agenda. It is interesting for less-resourced languages since it avoids the use of pronunciation dictionary, which is one of the major components in the traditional ASR systems. However, like any deep neural network (DNN) approaches, E2E is data greedy. This makes the application of E2E to less-resourced languages questionable. However, using data from other languages in a multilingual (ML) setup is being applied to solve the problem of data scarcity. We have, therefore, conducted ML E2E ASR experiments for four less-resourced Ethiopian languages using different language and acoustic modelling units. The results of our experiments show that relative Word Error Rate (WER) reductions (over the monolingual E2E systems) of up to 29.83% can be achieved by just using data of two related languages in E2E ASR system training. Moreover, we have also noticed that the use of data from less related languages also leads to E2E ASR performance improvement over the use of monolingual data.
In this paper, we describe an attempt towards the development of parallel corpora for English and... more In this paper, we describe an attempt towards the development of parallel corpora for English and Ethiopian Languages, such as Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. The corpora are used for conducting bi-directional SMT experiments. The BLEU scores of the bi-directional SMT systems show a promising result. The morphological richness of the Ethiopian languages has a great impact on the performance of SMT especially when the targets are Ethiopian languages.
In this paper, we describe an attempt towards the development of parallel corpora for English and... more In this paper, we describe an attempt towards the development of parallel corpora for English and Ethiopian Languages, such as Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. The corpora are used for conducting a bi-directional statistical machine translation experiments. The BLEU scores of the bi-directional Statistical Machine Translation (SMT) systems show a promising result. The morphological richness of the Ethiopian languages has a great impact on the performance of SMT specially when the targets are Ethiopian languages. Now we are working towards an optimal alignment for a bi-directional English-Ethiopian languages SMT.
In this paper, we present the cross-lingual and multilingual experiments we have conducted using ... more In this paper, we present the cross-lingual and multilingual experiments we have conducted using existing resources of other languages for the development of speech recognition system for less-resourced languages. In our experiments, we used the Globalphone corpus as source and considered four Ethiopian languages namely Amharic, Oromo, Tigrigna and Wolaytta as targets. We have developed multilingual (ML) Automatic Speech Recognition (ASR) systems and decoded speech of the four Ethiopian languages. A multilingual acoustic model (AM) trained with speech data of 22 Globalphone languages but the target languages, achieved a Word Error Rate (WER) of 15.79%. Moreover, including training speech of one closely related language (in terms of phonetic overlap) in ML AM training resulted in a relative WER reduction of 51.41%. Although adaptation of ML systems did not give significant WER reduction over the monolingual ones, it enables us to rapidly adapt existing ML ASR systems to new languages. In sum, our experiments demonstrated that ASR systems can be developed rapidly with a pronunciation dictionary (PD) of low out of vocabulary (OOV) rate and a strong language model (LM).
Development of Multilingual Automatic Speech Recognition (ASR) systems enables to share existing ... more Development of Multilingual Automatic Speech Recognition (ASR) systems enables to share existing speech and text corpora among languages. We have conducted experiments on the development of multilingual Acoustic Models (AM) and Language Models (LM) for Tigrigna. Using Amharic Deep Neural Network (DNN) AM, Tigrigna pronunciation dictionary and trigram LM, we achieved a Word Error Rate (WER) of 30.9% for Tigrigna. Adding training speech from the target language (Tigrigna) to the whole training speech of the donor language (Amharic) continuously reduces WER with the amount of added data. We have also developed different (including recurrent neural networks based) multilingual LMs and achieved a relative WER reduction of 3.56% compared to the use of monolingual trigram LMs. Considering scarcity of computational resources to decode with very large vocabularies, we have also experimented on the use of morphemes as pronunciation and language modeling units. We have achieved character error rate (CER) of 7.9% which is relatively lower by 38.3% to 1.3% than the CER of the word-based models of smaller vocabularies than 162k. Our results show the possibility of developing ASR system for an Ethio-Semitic language using an existing speech and text corpora of another language in the family.
This paper presents our investigation of the effect of language resources on the performance of A... more This paper presents our investigation of the effect of language resources on the performance of Amharic speech recognition. We have used language model training text of different sizes and seen the effect on word error rate (WER) reduction. Moreover, we have investigated the effect of handling language issues (germination, epenthetic vowel insertion and glottal stop consonant pronunciation) on the performance of speech recognition systems using data-driven phone-level transcriptions. The results of our experiments show that only slight reduction in WER can be obtained by increasing language model training text. However, proper transcription of gemination, the epenthetic vowel and the glottal stop consonant did not bring performance improvement for Amharic speech recognition. This can be attributed to the larger number of phone HMM acoustic models (62 compared to 37 phone set of the grapheme-based phone-level transcriptions) trained with a small (5 hrs) training speech.
This study investigated the possibility of developing a large vocabulary continuous speech recogn... more This study investigated the possibility of developing a large vocabulary continuous speech recognizer (LVCSR) for Amharic using the different syllable types V, CV, VC, CVC, VCC and CVCC found in the language as acoustic units. Syllables as longer length acoustic units are able to embed the spectral and temporal dependencies found in speech and thus able to model it well. The recognizer was developed using the Hidden Markov Model as a modeling technique. The result of the experiments shows that syllables are promising units for Amharic LVCSR provided that enough training data is available.
... Verbs are inflected for person, subject, ob-ject, gender, number and tense (Baye 2000 EC: 131)... more ... Verbs are inflected for person, subject, ob-ject, gender, number and tense (Baye 2000 EC: 131). ... The problems posed by Amharic morphology to language modeling were illus-trated by Solomon (2006) who, therefore, recommended the development of sub-word based ...
ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020
This paper describes new reference benchmark results based on hybrid Hidden Markov Model and Deep... more This paper describes new reference benchmark results based on hybrid Hidden Markov Model and Deep Neural Networks (HMM-DNN) for the GlobalPhone (GP) multilingual text and speech database. GP is a multilingual database of high-quality read speech with corresponding transcriptions and pronunciation dictionaries in more than 20 languages. Moreover, we provide new results for five additional languages, namely, Amharic, Oromo, Tigrigna, Wolaytta, and Uyghur. Across the 22 languages considered, the hybrid HMM-DNN models outperform the HMM-GMM based models regardless of the size of the training speech used. Overall, we achieved relative improvements that range from 7.14% to 59.43%.
ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020
In this work, we present speech recognition systems for four Ethiopian languages: Amharic, Tigrig... more In this work, we present speech recognition systems for four Ethiopian languages: Amharic, Tigrigna, Oromo and Wolaytta. We have used comparable training corpora of about 20 to 29 hours speech and evaluation speech of about 1 hour for each of the languages. For Amharic and Tigrigna, lexical and language models of different vocabulary size have been developed. For Oromo and Wolaytta, the training lexicons have been used for decoding. We achieved relative word error rate (WER) reductions for all the languages by using Deep Neural Networks (DNN) based acoustic models that range from 15.1% to 31.45%. The relative improvement obtained for Wolaytta speech recognition system is much higher (31.45%) than the improvement achieved for the other languages. This attributes to the weaker language model and the bigger size of training speech we used for Wolaytta.
In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian langua... more In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian languages (Amharic, Tigrigna, Oromo and Wolaytta). The aim of the analysis is to select speech data from GP for the development of multilingual Automatic Speech Recognition (ASR) system for the Ethiopian languages. To this end, phonetic overlaps among GP and Ethiopian languages have been analyzed. The result of our analysis shows that there is much phonetic overlap among Ethiopian languages although they are from three different language families. From GP, Turkish, Uyghur and Croatian are found to have much overlap with the Ethiopian languages. On the other hand, Korean has less phonetic overlap with the rest of the languages. Moreover, morphological complexity of the GP and Ethiopian languages, reflected by type to token ration (TTR) and out of vocabulary (OOV) rate, has been analyzed. Both metrics indicated the morphological complexity of the languages. Korean and Amharic have been identifie...
In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian langua... more In this paper, we present the analysis of GlobalPhone (GP) and speech corpora of Ethiopian languages (Amharic, Tigrigna, Oromo and Wolaytta). The aim of the analysis is to select speech data from GP for the development of multilingual Automatic Speech Recognition (ASR) system for the Ethiopian languages. To this end, phonetic overlaps among GP and Ethiopian languages have been analyzed. The result of our analysis shows that there is much phonetic overlap among Ethiopian languages although they are from three different language families. From GP, Turkish, Uyghur and Croatian are found to have much overlap with the Ethiopian languages. Moreover, morphological complexity of the GP and Ethiopian languages, reflected by type to token ration (TTR) and out of vocabulary (OOV) rate, has been analyzed. Korean and Amharic have been identified as extremely morphologically complex compared to the other languages. Tigrigna, Russian, Turkish and Polish are also among the morphologically complex l...
In this paper, we describe the development of parallel corpora for Ethiopian Languages: Amharic, ... more In this paper, we describe the development of parallel corpora for Ethiopian Languages: Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. To check the usability of all the corpora we conducted baseline bi-directional statistical machine translation (SMT) experiments for seven language pairs. The performance of the bi-directional SMT systems shows that all the corpora can be used for further investigations. We have also shown that the morphological complexity of the Ethio-Semitic languages has a negative impact on the performance of the SMT especially when they are target languages. Based on the results we obtained, we are currently working towards handling the morphological complexities to improve the performance of statistical machine translation among the Ethiopian languages. * This work is licenced under a Creative Commons Attribution 4.0 International License. Page numbers and proceedings footer are added by the organizers. License details: http://creativecommons.org/licenses/b...
It is known that Automatic Speech Recognition (ASR) is very useful for human-computer interaction... more It is known that Automatic Speech Recognition (ASR) is very useful for human-computer interaction in all the human languages. However, due to its requirement for a big speech corpus, which is very expensive, it has not been developed for most of the languages. Multilingual ASR (MLASR) has been suggested to share existing speech corpora among related languages to develop an ASR for languages which do not have the required speech corpora. Literature shows that phonetic relatedness goes across language families. We have, therefore, conducted experiments on MLASR taking two language families: one as source (Oromo from Cushitic) and the other as target (Wolaytta from Omotic). Using Oromo Deep Neural Network (DNN) based acoustic model, Wolaytta pronunciation dictionary and language model we have achieved Word Error Rate (WER) of 48.34% for Wolaytta. Moreover, our experiments show that adding only 30 minutes of speech data from the target language (Wolaytta) to the whole training data (22....
ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021
The End-to-End (E2E) approach, which maps a sequence of input features into a sequence of graphem... more The End-to-End (E2E) approach, which maps a sequence of input features into a sequence of graphemes or words, to Automatic Speech Recognition (ASR) is a hot research agenda. It is interesting for less-resourced languages since it avoids the use of pronunciation dictionary, which is one of the major components in the traditional ASR systems. However, like any deep neural network (DNN) approaches, E2E is data greedy. This makes the application of E2E to less-resourced languages questionable. However, using data from other languages in a multilingual (ML) setup is being applied to solve the problem of data scarcity. We have, therefore, conducted ML E2E ASR experiments for four less-resourced Ethiopian languages using different language and acoustic modelling units. The results of our experiments show that relative Word Error Rate (WER) reductions (over the monolingual E2E systems) of up to 29.83% can be achieved by just using data of two related languages in E2E ASR system training. Moreover, we have also noticed that the use of data from less related languages also leads to E2E ASR performance improvement over the use of monolingual data.
In this paper, we describe an attempt towards the development of parallel corpora for English and... more In this paper, we describe an attempt towards the development of parallel corpora for English and Ethiopian Languages, such as Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. The corpora are used for conducting bi-directional SMT experiments. The BLEU scores of the bi-directional SMT systems show a promising result. The morphological richness of the Ethiopian languages has a great impact on the performance of SMT especially when the targets are Ethiopian languages.
In this paper, we describe an attempt towards the development of parallel corpora for English and... more In this paper, we describe an attempt towards the development of parallel corpora for English and Ethiopian Languages, such as Amharic, Tigrigna, Afan-Oromo, Wolaytta and Ge’ez. The corpora are used for conducting a bi-directional statistical machine translation experiments. The BLEU scores of the bi-directional Statistical Machine Translation (SMT) systems show a promising result. The morphological richness of the Ethiopian languages has a great impact on the performance of SMT specially when the targets are Ethiopian languages. Now we are working towards an optimal alignment for a bi-directional English-Ethiopian languages SMT.
In this paper, we present the cross-lingual and multilingual experiments we have conducted using ... more In this paper, we present the cross-lingual and multilingual experiments we have conducted using existing resources of other languages for the development of speech recognition system for less-resourced languages. In our experiments, we used the Globalphone corpus as source and considered four Ethiopian languages namely Amharic, Oromo, Tigrigna and Wolaytta as targets. We have developed multilingual (ML) Automatic Speech Recognition (ASR) systems and decoded speech of the four Ethiopian languages. A multilingual acoustic model (AM) trained with speech data of 22 Globalphone languages but the target languages, achieved a Word Error Rate (WER) of 15.79%. Moreover, including training speech of one closely related language (in terms of phonetic overlap) in ML AM training resulted in a relative WER reduction of 51.41%. Although adaptation of ML systems did not give significant WER reduction over the monolingual ones, it enables us to rapidly adapt existing ML ASR systems to new languages. In sum, our experiments demonstrated that ASR systems can be developed rapidly with a pronunciation dictionary (PD) of low out of vocabulary (OOV) rate and a strong language model (LM).
Development of Multilingual Automatic Speech Recognition (ASR) systems enables to share existing ... more Development of Multilingual Automatic Speech Recognition (ASR) systems enables to share existing speech and text corpora among languages. We have conducted experiments on the development of multilingual Acoustic Models (AM) and Language Models (LM) for Tigrigna. Using Amharic Deep Neural Network (DNN) AM, Tigrigna pronunciation dictionary and trigram LM, we achieved a Word Error Rate (WER) of 30.9% for Tigrigna. Adding training speech from the target language (Tigrigna) to the whole training speech of the donor language (Amharic) continuously reduces WER with the amount of added data. We have also developed different (including recurrent neural networks based) multilingual LMs and achieved a relative WER reduction of 3.56% compared to the use of monolingual trigram LMs. Considering scarcity of computational resources to decode with very large vocabularies, we have also experimented on the use of morphemes as pronunciation and language modeling units. We have achieved character error rate (CER) of 7.9% which is relatively lower by 38.3% to 1.3% than the CER of the word-based models of smaller vocabularies than 162k. Our results show the possibility of developing ASR system for an Ethio-Semitic language using an existing speech and text corpora of another language in the family.
This paper presents our investigation of the effect of language resources on the performance of A... more This paper presents our investigation of the effect of language resources on the performance of Amharic speech recognition. We have used language model training text of different sizes and seen the effect on word error rate (WER) reduction. Moreover, we have investigated the effect of handling language issues (germination, epenthetic vowel insertion and glottal stop consonant pronunciation) on the performance of speech recognition systems using data-driven phone-level transcriptions. The results of our experiments show that only slight reduction in WER can be obtained by increasing language model training text. However, proper transcription of gemination, the epenthetic vowel and the glottal stop consonant did not bring performance improvement for Amharic speech recognition. This can be attributed to the larger number of phone HMM acoustic models (62 compared to 37 phone set of the grapheme-based phone-level transcriptions) trained with a small (5 hrs) training speech.
This study investigated the possibility of developing a large vocabulary continuous speech recogn... more This study investigated the possibility of developing a large vocabulary continuous speech recognizer (LVCSR) for Amharic using the different syllable types V, CV, VC, CVC, VCC and CVCC found in the language as acoustic units. Syllables as longer length acoustic units are able to embed the spectral and temporal dependencies found in speech and thus able to model it well. The recognizer was developed using the Hidden Markov Model as a modeling technique. The result of the experiments shows that syllables are promising units for Amharic LVCSR provided that enough training data is available.
... Verbs are inflected for person, subject, ob-ject, gender, number and tense (Baye 2000 EC: 131)... more ... Verbs are inflected for person, subject, ob-ject, gender, number and tense (Baye 2000 EC: 131). ... The problems posed by Amharic morphology to language modeling were illus-trated by Solomon (2006) who, therefore, recommended the development of sub-word based ...
Uploads
Papers by Martha Yifiru