An improved speech enhancement algorithm based on minimum mean square error of log-spectral amplitude estimator and masking property of human auditory system is proposed in this paper. The short-time spectral amplitude is estimated based on the minimum mean square error of log-spectral amplitude estimator. Residual musical noise can be masked by exploiting the masking properties of the human auditory system and a psycho acoustically motivated weighting filter is designed. The performance of our proposed algorithm is tested under various types of noise and at different SNR levels, and the results show that our proposed algorithm gives a superior performance as compared to the conventional algorithm and the residual musical noise is effectively masked.
{"title":"An Improved MMSE-LSA Speech Enhancement Algorithm Based on Human Auditory Masking Property","authors":"Yong Zhang, Yi Y. Liu","doi":"10.1109/IALP.2013.49","DOIUrl":"https://doi.org/10.1109/IALP.2013.49","url":null,"abstract":"An improved speech enhancement algorithm based on minimum mean square error of log-spectral amplitude estimator and masking property of human auditory system is proposed in this paper. The short-time spectral amplitude is estimated based on the minimum mean square error of log-spectral amplitude estimator. Residual musical noise can be masked by exploiting the masking properties of the human auditory system and a psycho acoustically motivated weighting filter is designed. The performance of our proposed algorithm is tested under various types of noise and at different SNR levels, and the results show that our proposed algorithm gives a superior performance as compared to the conventional algorithm and the residual musical noise is effectively masked.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133392891","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This work makes first attempt towards Chinese-Uyghur bilateral Machine Translation (MT) using an example based approach. We have developed a Chinese-Uyghur bilateral EBMT system by constructing bilingual corpus, implementing its storage format, developing similarity computation algorithm and recombining similar sentence. Finally, we evaluate our system by human judges, experimental result shows that there is still room for improvement for translation quality of our system.
{"title":"Implementation of Chinese-Uyghur Bilateral EBMT System","authors":"Kahaerjiang Abiderexiti, Tianfang Yao, Tuergen Yibulayin, Aishan Wumaier, Yasen Yiming","doi":"10.1109/IALP.2013.26","DOIUrl":"https://doi.org/10.1109/IALP.2013.26","url":null,"abstract":"This work makes first attempt towards Chinese-Uyghur bilateral Machine Translation (MT) using an example based approach. We have developed a Chinese-Uyghur bilateral EBMT system by constructing bilingual corpus, implementing its storage format, developing similarity computation algorithm and recombining similar sentence. Finally, we evaluate our system by human judges, experimental result shows that there is still room for improvement for translation quality of our system.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130172840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Eventive verb is a small class of Chinese verbs. Compared with other general verbs, it has different syntactic, semantic features. This article refers the verbs which can follow VP and must describe a complete event with the help of other verbs or the implicit verb as eventive verbs. Based on this, it gives the formal decision criteria of eventive verbs, and makes the quantitative and qualitative analysis of the internal connection and selective restrictions for the eventive verbs, which will be useful to set rules of the verb collocation.
{"title":"Judgment, Extraction and Selective Restriction of Chinese Eventive Verb","authors":"Mengxiang Wang, Houfeng Wang, Longkai Zhang","doi":"10.1109/IALP.2013.9","DOIUrl":"https://doi.org/10.1109/IALP.2013.9","url":null,"abstract":"Eventive verb is a small class of Chinese verbs. Compared with other general verbs, it has different syntactic, semantic features. This article refers the verbs which can follow VP and must describe a complete event with the help of other verbs or the implicit verb as eventive verbs. Based on this, it gives the formal decision criteria of eventive verbs, and makes the quantitative and qualitative analysis of the internal connection and selective restrictions for the eventive verbs, which will be useful to set rules of the verb collocation.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"2020 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114245929","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Chinese word segmentation (CWS) is an important and necessary problem to analyze Chinese texts. The state-of-art CWS systems are mostly based on sequence labeling algorithm and use the discriminative model with millions of overlapping binary features. However, there are few works on porting these systems to the devices with limited computing capacity and memory. In this paper, we focus on two challenges in Chinese word segmentation: (1) low accuracy of out-of-vocabulary word and (2) huge feature space. To resolve these two difficult problems, we propose a method to abstract the original input on both character and feature levels. We group the "similar'' features to generate more abstract representation. Experimental results show that feature abstraction can greatly reduce the feature space with a comparable performance.
{"title":"Feature Abstraction for Lightweight and Accurate Chinese Word Segmentation","authors":"Le Tian, Xipeng Qiu, Xuanjing Huang","doi":"10.1109/IALP.2013.65","DOIUrl":"https://doi.org/10.1109/IALP.2013.65","url":null,"abstract":"Chinese word segmentation (CWS) is an important and necessary problem to analyze Chinese texts. The state-of-art CWS systems are mostly based on sequence labeling algorithm and use the discriminative model with millions of overlapping binary features. However, there are few works on porting these systems to the devices with limited computing capacity and memory. In this paper, we focus on two challenges in Chinese word segmentation: (1) low accuracy of out-of-vocabulary word and (2) huge feature space. To resolve these two difficult problems, we propose a method to abstract the original input on both character and feature levels. We group the \"similar'' features to generate more abstract representation. Experimental results show that feature abstraction can greatly reduce the feature space with a comparable performance.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114522683","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As an agglutinative language, Uyghur has rich functional suffixes. This study is using functional suffixes (including suffix stringes) to define POS inventory for stem-suffix segmentation and word POS-Tagging based on Prof. Litip Tohti's studies on Uyghur generative syntax. To test the feasibility of functional suffix definition system, we conducted stem-suffix segmentation in a big text corpus (270,000 words, without repetition). About 14,100 most frequently used suffix strings generated. The limited number of suffix strings can reflect the feature of huge amount of Uyghur syntactic components. We tag the suffix strings as one unit based on Prof. Litip Tohti's study, and tag the words in corpus based on the suffix instead of the stem. This is a new try, we hope to improve the performance of Uyghur language processing.
{"title":"Uyghur Stem-Suffix Segmentation and POS-Tagging Based on Functional Suffixes","authors":"Haibo Wang, Yiqing Zu, Lang Wu, Litip Tohti","doi":"10.1109/IALP.2013.72","DOIUrl":"https://doi.org/10.1109/IALP.2013.72","url":null,"abstract":"As an agglutinative language, Uyghur has rich functional suffixes. This study is using functional suffixes (including suffix stringes) to define POS inventory for stem-suffix segmentation and word POS-Tagging based on Prof. Litip Tohti's studies on Uyghur generative syntax. To test the feasibility of functional suffix definition system, we conducted stem-suffix segmentation in a big text corpus (270,000 words, without repetition). About 14,100 most frequently used suffix strings generated. The limited number of suffix strings can reflect the feature of huge amount of Uyghur syntactic components. We tag the suffix strings as one unit based on Prof. Litip Tohti's study, and tag the words in corpus based on the suffix instead of the stem. This is a new try, we hope to improve the performance of Uyghur language processing.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116175811","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Speech signal feature extraction is an important part of speech recognition system. We present Tucker decomposition to extract speech feature. Firstly, the preprocessed speech signal is decomposed via three-level Wavelet transform, and the information in different scales is obtained. Next, the conventional feature parameters are extracted from the different scales, and a 3-order speech tensor (frames, scales, feature parameters) could be created. Then, the tensor is decomposed by Tucker decomposition, and projection matrices in different mode are obtained. Thirdly, matrix product is performed between speech tensor and projection matrices in each mode, and mapped results are metricized. Finally, feature system in high order space is built, in other words, speech feature matrices are obtained. The feature system can fully express speech signal features. These matrices can be used for model training and speech recognition. Numerical experiments support the advantage of Tucker decomposition over conventional methods for speech signal feature extraction, furthermore, it is robust to noisy speech.
{"title":"Application of Tucker Decomposition in Speech Signal Feature Extraction","authors":"Lidong Yang, Jing Wang, Xiang Xie, Jingming Kuang","doi":"10.1109/IALP.2013.50","DOIUrl":"https://doi.org/10.1109/IALP.2013.50","url":null,"abstract":"Speech signal feature extraction is an important part of speech recognition system. We present Tucker decomposition to extract speech feature. Firstly, the preprocessed speech signal is decomposed via three-level Wavelet transform, and the information in different scales is obtained. Next, the conventional feature parameters are extracted from the different scales, and a 3-order speech tensor (frames, scales, feature parameters) could be created. Then, the tensor is decomposed by Tucker decomposition, and projection matrices in different mode are obtained. Thirdly, matrix product is performed between speech tensor and projection matrices in each mode, and mapped results are metricized. Finally, feature system in high order space is built, in other words, speech feature matrices are obtained. The feature system can fully express speech signal features. These matrices can be used for model training and speech recognition. Numerical experiments support the advantage of Tucker decomposition over conventional methods for speech signal feature extraction, furthermore, it is robust to noisy speech.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128240673","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lexical semantic network is one of the important ways to express lexical semantics, which plays an important role in semantic analysis, semantic disambiguation, semantic reasoning and semantic summarization etc. With the deep development of semantic processing, the research work of Mongolian information processing highly needs the support of lexical semantic network. This research referred to the construction theory of Chinese and English Word Net, proposing a frame designing method of Mongolian noun semantic network which is compatible with Word Net, and basically realized the presentation of Mongolian synonym sets and the construction of semantic relations between these sets. This article mainly introduced the construction method of the semantic relations in the Mongolian noun semantic network.
{"title":"The Frame Design of Mongolian Noun Semantic Network","authors":"Hasi, Enbo Tang","doi":"10.1109/IALP.2013.73","DOIUrl":"https://doi.org/10.1109/IALP.2013.73","url":null,"abstract":"Lexical semantic network is one of the important ways to express lexical semantics, which plays an important role in semantic analysis, semantic disambiguation, semantic reasoning and semantic summarization etc. With the deep development of semantic processing, the research work of Mongolian information processing highly needs the support of lexical semantic network. This research referred to the construction theory of Chinese and English Word Net, proposing a frame designing method of Mongolian noun semantic network which is compatible with Word Net, and basically realized the presentation of Mongolian synonym sets and the construction of semantic relations between these sets. This article mainly introduced the construction method of the semantic relations in the Mongolian noun semantic network.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125716159","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To get better performance, Some researchers have proposed relative work to exploit the position and proximity information of query terms in language model. However these models need large quantity of training data and its computation complexity is comparatively high. This paper presents an information retrieval model combining sentence level retrieval and use sentence as a unit to compute the relevant degree of the sentence to query. Experiment results show our model can get better performance than baseline models.
{"title":"Information Retrieval Model Combining Sentence Level Retrieval","authors":"Jiali Zuo, Mingwen Wang, Jianyi Wan, Wenbing Luo","doi":"10.1109/IALP.2013.76","DOIUrl":"https://doi.org/10.1109/IALP.2013.76","url":null,"abstract":"To get better performance, Some researchers have proposed relative work to exploit the position and proximity information of query terms in language model. However these models need large quantity of training data and its computation complexity is comparatively high. This paper presents an information retrieval model combining sentence level retrieval and use sentence as a unit to compute the relevant degree of the sentence to query. Experiment results show our model can get better performance than baseline models.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125227031","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Spam filter benefits from two state-of-the-art discriminative models: Logistic Regression (LR) and Relaxed Online Support Vector Machine (ROSVM). It is natural that two models reach their optimal performance after different training examples. We presented a combination model which integrated LR and ROSVM into a unified one. We divided the training process into two phases. In the first phase, LR was used as filtering model to train and learn, at the same time ROSVM accepted the right result to learn. In the second phase, ROSVM was used as filtering model to train after a point which was found in experiments. Experimental results on the public data sets (TREC06-c, TREC06-p, TREC07-p) showed that the combination of ROSVM and LR spam filter gave the better performance than LR filter and ROSVM filter in immediate feedback.
{"title":"Combination of ROSVM and LR for Spam Filter","authors":"Yadong Wang, Haoliang Qi, Hong Deng, Yong Han","doi":"10.1109/IALP.2013.60","DOIUrl":"https://doi.org/10.1109/IALP.2013.60","url":null,"abstract":"Spam filter benefits from two state-of-the-art discriminative models: Logistic Regression (LR) and Relaxed Online Support Vector Machine (ROSVM). It is natural that two models reach their optimal performance after different training examples. We presented a combination model which integrated LR and ROSVM into a unified one. We divided the training process into two phases. In the first phase, LR was used as filtering model to train and learn, at the same time ROSVM accepted the right result to learn. In the second phase, ROSVM was used as filtering model to train after a point which was found in experiments. Experimental results on the public data sets (TREC06-c, TREC06-p, TREC07-p) showed that the combination of ROSVM and LR spam filter gave the better performance than LR filter and ROSVM filter in immediate feedback.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122474471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
By comparing the Eye Movement on data of Japanese students reading Chinese texts with or without marks for word boundaries, the researcher has found that: (1) To Japanese students, the Chinese words represent more Psychological reality than Chinese characters. (2) The reading efficiency of Japanese students could be immensely improved by inserting word boundary marks. It is, therefore, suggested that when publishing Chinese textbooks intended for Japanese students, we adjust the way we do typesetting today and insert word boundary marks into the text.
{"title":"A Study on Japanese Students (Intermediate Level) Reading Chinese Texts With or Without Marks for Word Boundaries","authors":"Yu Peng","doi":"10.1109/IALP.2013.38","DOIUrl":"https://doi.org/10.1109/IALP.2013.38","url":null,"abstract":"By comparing the Eye Movement on data of Japanese students reading Chinese texts with or without marks for word boundaries, the researcher has found that: (1) To Japanese students, the Chinese words represent more Psychological reality than Chinese characters. (2) The reading efficiency of Japanese students could be immensely improved by inserting word boundary marks. It is, therefore, suggested that when publishing Chinese textbooks intended for Japanese students, we adjust the way we do typesetting today and insert word boundary marks into the text.","PeriodicalId":413833,"journal":{"name":"2013 International Conference on Asian Language Processing","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123340271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：