检索结果-内蒙古大学图书馆

您好，读者！请登录

内蒙古大学图书馆

首页
概况
党建
资源
服务
科研支持
- 论文收录引用证明
- 科技查新
知识产权
档案馆
帮助

咨询与建议

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

您的常用邮箱：*

您的手机号码：*

问题描述：

当前已输入0个字，您还可以输入200个字

全部搜索
期刊论文
图书
学位论文
标准
纸本馆藏
外文资源发现
数据库导航
超星发现

高级检索

时间限定

出版年份：

文献类型

图书期刊文献学位论文多媒体

馆藏选择

电子馆藏纸本馆藏

核心期刊

全部期刊 SCI 收录期刊 SSCI 收录期刊 EI 收录期刊 CSCD 收录期刊 CSSCI 收录期刊

语言

中文英文

文献类型

期刊文献图书学位论文标准纸本馆藏

帮助

文字说明：

T=题名（书名、题名），A=作者（责任者），K=主题词，P=出版物名称，PU=出版社名称，O=机构（作者单位、学位授予单位、专利申请人），L=中图分类号，C=学科分类号，U=全部字段，Y=年（出版发行年、学位年度、标准发布年）

检索规则说明：

AND代表“并且”；OR代表“或者”；NOT代表“不包含”；(注意必须大写,运算符两边需空一格)

检索范例：

范例一：(K=图书馆学 OR K=情报学) AND A=范并思 AND Y=1982-2016
范例二：P=计算机应用与软件 AND (U=C++ OR U=Basic) NOT K=Visual AND Y=2011-2016

分类表

所选分类

>> <<

限定检索结果

文献类型

139 篇 会议
68 篇 期刊文献
1 册 图书

馆藏范围

208 篇 电子文献
0 种 纸本馆藏

日期分布

学科分类号

136 篇 工学
- 68 篇 计算机科学与技术...
- 61 篇 软件工程
- 42 篇 信息与通信工程
- 16 篇 生物医学工程（可授...
- 15 篇 生物工程
- 14 篇 光学工程
- 13 篇 电气工程
- 11 篇 电子科学与技术（可...
- 10 篇 控制科学与工程
- 7 篇 仪器科学与技术
- 4 篇 化学工程与技术
- 3 篇 建筑学
- 3 篇 土木工程
93 篇 理学
- 59 篇 物理学
- 26 篇 数学
- 19 篇 生物学
- 13 篇 统计学（可授理学、...
- 12 篇 系统科学
- 5 篇 化学
24 篇 管理学
- 14 篇 图书情报与档案管...
- 8 篇 管理科学与工程(可...
22 篇 医学
- 20 篇 临床医学
- 14 篇 基础医学(可授医学...
- 6 篇 公共卫生与预防医...
- 5 篇 药学(可授医学、理...
6 篇 文学
- 4 篇 外国语言文学
- 3 篇 中国语言文学
5 篇 法学
- 5 篇 社会学
3 篇 经济学
- 3 篇 应用经济学
3 篇 教育学
- 3 篇 心理学(可授教育学...
2 篇 农学
1 篇 哲学
1 篇 历史学
1 篇 艺术学

主题

32 篇 speech recogniti...
23 篇 hidden markov mo...
15 篇 humans
14 篇 natural language...
11 篇 speech
10 篇 semantics
8 篇 training data
8 篇 decoding
7 篇 error analysis
6 篇 support vector m...
6 篇 automatic speech...
6 篇 training
5 篇 speech processin...
5 篇 acoustics
5 篇 feature extracti...
5 篇 testing
4 篇 machine translat...
4 篇 loudspeakers
4 篇 data mining
4 篇 computational mo...

机构

11 篇 hkust human lang...
8 篇 human language t...
7 篇 human language t...
6 篇 human language t...
5 篇 university colle...
5 篇 human language t...
5 篇 human language t...
4 篇 leeds institute ...
4 篇 center for psych...
4 篇 department of in...
4 篇 institute for pa...
4 篇 department of st...
4 篇 university of pi...
4 篇 arnie charbonnea...
4 篇 ilsbio llc bioba...
4 篇 department of mo...
4 篇 department of bi...
4 篇 faculty of biosc...
4 篇 osaka internatio...
4 篇 tri-institutiona...

作者

35 篇 fung pascale
25 篇 pascale fung
20 篇 wu dekai
7 篇 justin jian zhan...
6 篇 bertero dario
6 篇 saers markus
6 篇 lo chi-kiu
6 篇 liu yi
5 篇 qu jianan y.
5 篇 khudanpur sanjee...
4 篇 he sicong
4 篇 hermansky hynek
4 篇 wang chengyan
4 篇 liu wai kat
4 篇 yang guang
4 篇 dekai wu
4 篇 qu xiaobo
4 篇 ganapathy sriram
3 篇 duan lixin
3 篇 shi zhang

语言

199 篇 英文
6 篇 其他
3 篇 中文

检索条件"机构=Human Language Technology Center Department of Electronic and Computer Engineering"

共 208 条记录，以下是191-200 订阅

全选清除本页清除全部题录导出标记到"检索档案"

详细简洁

排序：

Soft GPD for minimum classification error rate training 25

Soft GPD for minimum classification error rate training

引用

25th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2000

作者： Shi, Bertram E. Yao, Kaisheng Cao, Zhigang Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Kowloon Hong Kong Department of Electronic Engineering Tsinghua University Beijing100084 China

ISBN: (纸本)0780362934

Minimum classification error (MCE) rate training is a discriminative training method which seeks to minimize an empirical estimate of the error probability derived over a training set. The segmental generalized probabilistic descent (GPD) algorithm for MCE uses the log likelihood of the best path as a discriminant function to estimate the error probability. This paper shows that by using a discriminant function similar to the auxiliary function used in EM, we can obtain a «soft» version of GPD in the sense that information about all possible paths is retained. Complexity is similar to segmental GPD. For certain parameter values, the algorithm is equivalent to segmental GPD. By modifying the misclassification measure usually used, we can obtain an algorithm for embedded MCE training for continuous speech which does not require a separate N-best search to determine competing classes. Experimental results show error rate reduction of 20% compared with maximum likelihood training. © 2000 IEEE.

关键词： Maximum likelihood

来源：评论

学校读者我要写书评

暂无评论

An LLR-based technique of frame selection for GMM-based text-independent speaker identification 6

An LLR-based technique of frame selection for GMM-based text...

引用

6th International Conference on Spoken language Processing, ICSLP 2000

作者： Kucn, Tsoi Pang Pascalc, Fung Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong Hong Kong

ISBN: (纸本)7801501144

In speaker recognition systems, frame selection, which aims at determining which frame is useful and which is not and selecting useful frames from the test utterance, can be utilized to increase recognition accuracy. In this paper, we present a new approach for frame selection using Log Likelihood Ratio (LLR), which is based on the idea that if a frame contains speaker information, the Log likelihood Score of the corresponding speaker model will be much larger than that of its competing model. As a result, for each frame we can calculate the Log Likelihood Ratio (LLR) between the largest score and the second largest score in different speaker models and take it as a reference: Those frames with a small LLR can be rejected and those with a large LLR can be kept. This algorithm is implemented based on a GMM-based text-independent speaker identification system. We compare the algorithm with another frame selection approach based on Jensen Difference (JD). Experiment shows that the approach using JD reduces the error by about 39.34%, while our approach using LLR reduces the error by about 46.32%.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

Residual noise compensation for robust speech recognition in nonstationary noise 25

Residual noise compensation for robust speech recognition in...

引用

25th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2000

作者： Yao, Kaisheng Shi, Bertram E. Fung, Pascale Cao, Zhigang Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong Hong Kong Department of Electronic Engineering Tsinghua University Beijing100084 China

ISBN: (纸本)0780362934

We present a model-based noise compensation algorithm for robust speech recognition in nonstationary noisy environments. The effect of noise is split into a stationary part, compensated by parallel model combination, and a time varying residual. The evolution of residual noise parameters is represented by a set of state space models. The state space models are updated by Kalman prediction and the sequential maximum likelihood algorithm. Prediction of residual noise parameters from different mixtures are fused, and the fused noise parameters are used to modify the linearized likelihood score of each mixture. Noise compensation proceeds in parallel with recognition. Experimental results demonstrate that the proposed algorithm improves recognition performance in highly nonstationary environments, compared with parallel model combination alone. © 2000 IEEE.

关键词： Mixtures

来源：评论

学校读者我要写书评

暂无评论

MLLR-based accent model adaptation without accented data 6

MLLR-based accent model adaptation without accented data

引用

6th International Conference on Spoken language Processing, ICSLP 2000

作者： Liu, Wai Kat Fung, Pascale Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong China Weniwen Inc. United States

ISBN: (纸本)7801501144

When the user has an accent different from what the automatic speech recognition system is trained with, the performance of the systems degrades. This is attributed to both acoustic and phonological differences between accents. The phonological differences between two accents are due to different phoneme inventories in two languages. Even for the same phoneme, foreigners and native speakers pronounce different sounds. Since accented data is rare but monolingual data is abundant we propose using the accented speaker' s first language data directly instead of accented data in the second language for our purpose. We propose adapting the native English phoneme models to accented phoneme models using first language data in MLLR adaptation. The baseline performance is 35.25% (phone accuracy) in using native English phone models to recognize Cantoneseaccented English speech data. We compare accent adaptation by using accented data and source language data. On the average, using accented data for adaptation improves the phone accuracy by 69.98% while using source language data for adaptation improves the phone accuracy by 70.13%. This shows that both kinds of adaptation data give similar improvements. Therefore non-accented data can be used for adaptation. We can rapidly obtain an accent-adapted acoustic model without the need of collecting accented database.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

An LLR-Based Technique of Frame Selection for GMM-Based Text-Independent Speaker Identification

An LLR-Based Technique of Frame Selection for GMM-Based Text...

引用

6～(th) International Conference on Spoken language Processing

作者： TSOI PangKucn FUNG Pascale Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Clear Water Bay Hong Kong

In speaker recognition systems, frame selection, which aims at determining which frame is useful and which is not and selecting useful frames from the test utterance, can be utilized to increase recognition accuracy. In this paper, we present a new approach for frame selection using Log Likelihood Ratio (LLR), which is based on the idea that if a frame contains speaker information, the Log likelihood Score of the corresponding speaker model will be much larger than that of its competing model. As a result, for cach frame we can calculate the Log Likelihood Ratio (LLR) between the largest score and the second largest score in different speaker models and take it as a reference: Those frames with a small LLR can be rejected and those with a large LLR can be kept. This algorithm is implemented based on a GMM-hased text-independent speaker identification system. We compare the algorithm with another frame selection approach based on Jensen Difference (JD). Experiment shows that the approach using JD reduces the error by about 39.34％, while our approach using LLR reduces the error by about 46.32％.

关键词： LLR GMM An LLR-Based Technique of Frame Selection for GMM-Based Text-Independent Speaker Identification

来源：评论

学校读者我要写书评

暂无评论

PRINCIPAL MIXTURE SPEAKER ADAPTATION FOR IMPROVED CONTINUOUS SPEECH RECOGNITION

PRINCIPAL MIXTURE SPEAKER ADAPTATION FOR IMPROVED CONTINUOUS...

引用

6～(th) International Conference on Spoken language Processing

作者： PascaleFUNG Human Language Technology Center (HLTC Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology (HKUST

Nowadays, almost all speaker-independent (SI) speech recognition systems use CDHMM with multivariate mixture Gaussian as observation density to cover speaker variabilities. It has been shown that given sufficient training data, the more mixtures are used in the HMM observation density, the better the system's perform. However, acoustic HMM with more Gaussian densities is more complex and slows down recognition speed. Another efficient way to handle speaker variation is to use speaker

关键词： PRINCIPAL MIXTURE SPEAKER ADAPTATION FOR IMPROVED CONTINUOUS SPEECH RECOGNITION

来源：评论

学校读者我要写书评

暂无评论

NLLR-based accent model adaptation without accented data

NLLR-based accent model adaptation without accented data

引用

6～(th) International Conference on Spoken language Processing

作者： LIU Wai Kat Pascale FUNG Human Language Technology Center Department of Electrical and Electronic Engineering University of Science and Technolgy Clear Water Bay Hong Kong Human Language Technology Center Department of Electrical and Electronic Engineering University of Science and Technolgy Clear Water Bay Hong Kong

When the user has an accent different from what the automatic speech recognization system is trained with, the performance of the systems degrades. This is attributed to both acoustic and phonological differences between accents. The phonological differences between two accents are due to different phoneme inventories in two languages. Even for the same phoneme, foreigners and native speukers pronounce different sounds. Since accented data is rare but monolingual data is abundant, we propose using the accented speaker's first language data directly instead of accented data in the second language for our purpose. We propose adapting the native English phoneme models to accented phoneme models using first language data in MLLR adaptation. The baseline performance is 35.25% (phone accuracy) in using native English phone models to recognize Cantonese-accented English speech data. We compare accent adaptation by using accented data and source language data. On the average, using accented data for adaptation improves the phone accuracy by 69.98% while using source language data for adaptation improves the phone accuracy by 70.13%. This shows that both kinds of adaptation data give similar improvements. Therefore non-accented data call be used for adaptation. We can rapidly obtain an accent-adapted acoustic model without the need of collecting accented database.

关键词： NLLR-based accent model adaptation without accented data

来源：评论

学校读者我要写书评

暂无评论

Residual noise compensation for robust speech recognition in nonstationary noise

Residual noise compensation for robust speech recognition in...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Kaisheng Yao B.E. Shi Pascale Fung Zhigang Cao Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Hong Kong China

关键词： Noise robustness Speech recognition Working environment noise Additive noise State-space methods Hidden Markov models Speech enhancement Statistics Kalman filters Maximum likelihood estimation

来源：评论

学校读者我要写书评

暂无评论

Soft GPD for minimum classification error rate training

Soft GPD for minimum classification error rate training

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： B.E. Shi Kaisheng Yao Zhigang Cao Human Language Technology Center Department of Electrical and Electronic Engineering Hong Kong University of Science and Technology Kowloon Hong Kong China Department of Electronic Engineering Tsinghua University Beijing China

Minimum classification error (MCE) rate training is a discriminative training method which seeks to minimize an empirical estimate of the error probability derived over a training set. The segmental generalized probabilistic descent (GPD) algorithm for MCE uses the log likelihood of the best path as a discriminant function to estimate the error probability. This paper shows that by using a discriminant function similar to the auxiliary function used in EM, we can obtain a "soft" version of GPD in the sense that information about all possible paths is retained. Complexity is similar to segmental GPD. For certain parameter values, the algorithm is equivalent to segmental GPD. By modifying the misclassification measure usually used, we can obtain an algorithm for embedded MCE training for continuous speech which does not require a separate N-best search to determine competing classes. Experimental results show error rate reduction of 20% compared with maximum likelihood training.

关键词： Error analysis Hidden Markov models Error probability Speech Maximum likelihood estimation Weight control Loss measurement

来源：评论

学校读者我要写书评

暂无评论

MAP-BASED CROSS-language ADAPTATION AUGMENTED BY LINGUISTIC KNOWLEDGE: FROM ENGLISH TO CHINESE 6

MAP-BASED CROSS-LANGUAGE ADAPTATION AUGMENTED BY LINGUISTIC ...

引用

6th European Conference on Speech Communication and technology, EUROSPEECH 1999

作者： Fung, Pascale Ma, Chi Yuen Liu, Wai Kat Human Language Technology Center Department of Electrical and Electronic Engineering University of Science and Technology HKUST Clear Water Bay Hong Kong

Construction of a recognizer in a new target language usually involves collection of a comprehensive database in that language as well as manual annotation and model training. For rapid development of new language recognizer, we propose (1) substituting target language phoneme models by source language phoneme models trained previously;and (2) adapting target language phoneme models from source language phoneme models using maximum a posteriori method. Evaluations show that adaptation method results in an error reduction rate of 51% from using substitution, and an error reduction rate of 10% from language-dependent training. We also propose using data-driven method with linguistic knowledge, instead of heuristic rules, to align source language phonemes to target language phonemes, both for substitution and adaptation. Evalution shows that using linguistic rules brings a 2% increase in recognition rate. © 1999 6th European Conference on Speech Communication and technology, EUROSPEECH 1999. All rights reserved.

关键词： Speech communication

来源：评论

学校读者我要写书评

暂无评论

没有更多数据了...

全选清除本页清除全部题录导出标记到“检索档案”

共21页 << < 12 13 14 15 16 17 18 19 20 21 > >>

检索报告对象比较合并检索0

隐藏清空

合并搜索

回到顶部

执行限定条件

内容：

评分：

请选择保存的检索档案：

请选择收藏分类：

订阅名称：

通借通还

温馨提示：

图书名称：

借书校区：

取书校区：

手机号码：

邮箱地址：

一卡通帐号：

电话和邮箱必须正确填写，我们会与您联系确认。

联系人：

所在院系：

联系邮箱：

联系电话：

内蒙古自治区呼和浩特市赛罕区大学西街235号邮编: 010021

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：