位置：成果数据库 > 期刊 > 期刊详情页
特定领域的汉语语言模型平滑算法比较研究
  • ISSN号：1002-8331
  • 期刊名称：《计算机工程与应用》
  • 时间：0
  • 分类：TP391.4[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]
  • 作者机构：[1]中科院声学所,中科信利语音实验室,北京100080
  • 相关基金：国家自然科学基金资助项目（编号：60535030）.
作者： 杨琳[1], 张建平[1], 颜永红[1]
关键词： 语言模型, 特定领域, 语音识别, 平滑, 字典, language model,specific domain,speech recognition,smoothing algorithm,dictionary
中文摘要：

为了完成特定领域的语音识别任务,利用有限的语料建立高性能的语言模型成为提高系统性能的关键.针对此问题,对特定领域的语言模型进行了研究.提出了利用高频新词来加强模型的领域特征的方法,采取了两种方案：一种是将高频新词直接加入原有字典,并在训练过程中增加这些新词的权重,使模型更能表达与领域相关的特征;一种是基于高频新词统计出一个和领域相关的小词表,并对这两种方案进行了比较研究.通过实验研究了适合汉语语言的平滑策略.最后,实验结果表明,对于特定领域问题,语言模型平滑算法对模型性能影响较大;采用适合汉语的Witten-Bell插值平滑,可以使识别率达到88.4%,比通用模型性能相对提高了18.18%.

英文摘要：

It is important to build a powerful language model by using limited corpora in the field of speech recognition for a specific domain.To deal with this problem,two methods concerning how to process new words with high frequencies in a specific domain are presented.One way is to add the new words to the dictionary directly and then give them a high weight in the procedure of training.The other is to work out a new dictionary according to the new words. And based on some comparative experiments,these two methods and various smoothing algorithms are studied in detail. At last,it can be concluded that the performance of language model is affected by the smoothing algorithm greatly,and the Witten-Bell interpolation method could improve the recognition rate to 88.4%,which is 18.18% higher than the general language model.

同期刊论文项目
汉语与英语听感知差异机理及应用研究
期刊论文 88 会议论文 112 著作 1
同项目期刊论文
细菌性脑膜炎致聋研究进展
外侧杏仁核在恐惧性条件化学习和记忆中的作用
基于网络的无监督MLLR?自适应算法实现及改进
Optimal?Time-Reversal?Focusing?by?an?Iterative?Least?Squares?Method
Perceptual?integration?between?target?speech?and?target-speech?reflection?reduces?masking?for?target
Both?frequency?and?interaural?delay?affect?ERP?responses?to?binaural?gap
An LVCSR Based Reading Miscue Detection System Using Knowledge of Reference
Automatic Singing Performance Evaluation for Untrained Singers
A two-element-microphone-array-based speech recognition system in vehicle environment
Discrimination between pathological normal voices using GMM-SVM approach
基于频率调制信息的人工耳蜗语音处理算法研究
语言声学进展及其应用
Precedence?effect-induced?enhancement?of?prepulse?inhibition?in?socially?reared?but?not?isolation-re
Auditory?frequency-following?responses?in?rat?ipsilateral?inferior?colliculus
The?dual-pathway?model?of?auditory?signal?processing
A?robust?real-time?decoder?using?memory-efficient?state?network
Automatic?language?identification?with?discriminative?language?characterization?based?on?SVM
Using SVM as Back-End Classifier for Language Identification
Robust?Speaker?Clustering?Using?Affinity?Propagation
Effective?Acoustic?Modeling?for?Pronunciation?Quality?Scoring?of?Strongly?Accented?Mandarin?Speech
Melody?Track?Selection?Using?Discriminative?Language?Model
The?effect?of?voice?cuing?on?releasing?Chinese?speech?from?informational?masking
The?experimental?investigation?on?the?reproduction?of?audible?sound?from?two?ultrasonic?beams
Effect?of?the?temporal?fine?structure?in?different?frequency?bands?on?Mandarin?tone?perception
Chronic?administration?of?clozapine?alleviates?reversal?learning?impairment?in?isolation?reared?rats
A Fast Calculation Method for the Scattered Sound Fields from Two Sound Beams Using the Gaussian-Bea
Generation of Audible Sound from Two Ultrasonic Beams with Dummy Head
Speech Enhancement Using Compact Microphone Array Applications in Distant Speech Acquisition
Development of a Mandarin-English Biligual Speech Recognition System with Unified Acoustic Models
Estimation?of?ICTD?in?Frequency?Sub-bands?Based?on?NDFT
Development?of?a?Mandarin-English?Bilingual?Speech?Recognition?System?for?Real?World?Music?Retrieval
“鸡尾酒会”环境中的知觉线索的去掩蔽作用
基于混合模型状态修正算法的非母语语音识别
嵌入式语音识别中一种高效的图搜索算法
基于听觉感知子带的自适应谷点形成算法
Fast?fuzzy?keyword?spotting?using?syllable?confusion?network
一种改进的基于层次聚类的说话人自动聚类算法
窄带的自同步音频水印算法
Ultrasonic Intruder Detection System for Home Security
Orthogonal Relief Algorithm for Feature Selection
The?influence?of?the?perceptual?or?fear?learning?on?rats’?prepulse?inhibition?induced?by?changes?in?
Distance-dependent Head-related Transfer Functions Measured with High Spatial Resolution Using a Spa
Emotional learning enhances stimulus-specific top-down modulation of sensorimotor gating in socially
Top-down modulation of prepulse inhibition the startle reflex in humans and rats
Detection of the break in interaural correlation is affected by interaural delay, aging, and center
Product HMM based training method for acoustic model with multiple size units
Target-Oriented?Acoustic?Radiation?Generation?Technique?for?Sound?Field?Control
The investigation of localized sound generation using two ultrasound beams
Speech?endpoint?detection?in?real?noise?environments
Auditory?fear?conditioning?modulates?prepulse?inhibition?in?socially-reared?rats?and?isolation-reare
Speech?Enhancement?Using?Improved?Adaptive?Null-?Forming?in?Frequency?Domain?with?Postfilter
Rapid?Adaptation?Based?on?Regression?Analysis
复杂噪声环境中的语音端点检测
一种任务域无关的语音关键词检测系统
Metabotropic?glutamate?subtype?5?receptors?modulate?fear-conditioning?induced?enhancement?of?prepuls
Auditory?evoked?responses?in?the?rat:?transverse?subdermal?electrodes?register?before?cochlear?nucle
Ultrasonic?auditory?evoked?response?recorded?in?the?rat’s?cochlear?nucleus
Human?Face?Classification?Using?Ultrasonic?Sonar?Imaging
Binaural?unmasking?of?frequency-following?responses?in?rat?amygdale
Product HMM-based training method for acoustic model with multiple-size units
The First International Symposium on Neurobehavioral Science in China (Editorial)
Simulated?phase-locking?stimulation:?an?improved?speech?processing?strategy?for?cochlear?implant
An?improved?cochlear?implant?strategy?incorporating?frequency?modulation?information
Using A Kind of Novel Phonotactic Information for SVM based Speaker Recognition
Metabotropic?glutamate?receptors?subtype?5?are?necessary?for?the?enhancement?of?auditory?evoked?pote
Rapid?calculations?of?the?scattered?sound?fields?generated?by?two?sound?beams
音频信号截幅失真的检测与修复
交叉对数似然度和贝叶斯信息判据的说话人聚类算法
基于人耳听觉模型的自动嗓音评估方法
汉语自然口语中声调识别的研究
中英双语混合语音识别研究
基于概念网络的短文本分类方法
汉语、英语听感知差异及适合汉语的人工耳蜗编码策略
一种基于联合源-滤波器模型优化的语音声门源模型估计方法
Cochlear implant signal processing algorithm based on frequency modulation
期刊信息
  • 《计算机工程与应用》
  • 北大核心期刊（2014版）
  • 主管单位:中国电子科技集团公司
  • 主办单位:华北计算技术研究所
  • 主编：怀进鹏
  • 地址：北京市海淀区北四环中路211号北京619信箱26分箱
  • 邮编：100083
  • 邮箱：ceaj@vip.163.com
  • 电话：
  • 国际标准刊号：ISSN：1002-8331
  • 国内统一刊号：ISSN：11-2127/TP
  • 邮发代号:82-605
  • 获奖情况:
  • 1. 2012年首批获得中国学术文献评价中心发布的 “...,2. 2001年获得新闻出版署“中国期刊方阵双效期刊”,3. 2008年首批入选国家科技部“中国精品科技期刊...,4.2003年-2011年连续获得工业和信息化部期刊最高...
  • 国内外数据库收录:
  • 俄罗斯文摘杂志,波兰哥白尼索引,美国剑桥科学文摘,英国科学文摘数据库,日本日本科学技术振兴机构数据库,中国中国科技核心期刊,中国北大核心期刊（2004版）,中国北大核心期刊（2008版）,中国北大核心期刊（2014版）,中国北大核心期刊（2000版）
  • 被引量:97887