检索结果-南通市图书馆

Integrated Chinese Segmentation, Parsing and Named Entity Recognition

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2018年第4期27卷 756-760页

作者： LI Dongchen ZHANG Xiantao wu xihong Key Laboratory of Machine Perception and Intelligence Speech and Hearing Research Center Peking University

Segmentation, named entity recognition and parsing are standalone techniques in natural language processing community, and their annotations are inconsistent. However, the joint output is needed in some practical use, and they rely on the result of each other to make more concise output. A unified model is learned to resolve these three tasks simultaneously. At the training stage, the joint annotation of the three tasks are employed to learn a unified model. At the decoding stage, the three tasks are carried out on a given text to provide a consistent *** results demonstrate the higher performance for each task and verify the benefits of the unified framework.

关键词： Parsing Joint representation Named entity recognition

Audio Mixing Inversion via Embodied Self-supervised Learning

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Machine Intelligence Research 2024年第1期21卷 55-62页

作者： Haotian Zhou Feng Yu xihong wu Department of AI Music and Music Information Technology Central Conservatory of MusicBeijing 100031China Laboratory of Music Artificial Intelligence Laboratory of Philosophy and Social SciencesMinistry of EducationBeijing 100031China School of Intelligence Science and Technology Peking UniversityBeijing 100871China

Audio mixing is a crucial part of music *** analyzing or recreating audio mixing,it is of great importance to conduct research on estimating mixing parameters used to create mixdowns from music recordings,i.e.,audio mixing ***,approaches of audio mixing inversion are rarely explored.A method of estimating mixing parameters from raw tracks and a stereo mixdown via embodied self-supervised learning is *** this work,several commonly used audio effects including gain,pan,equalization,reverb,and compression,are taken into *** method is able to learn an inference neural network that takes a stereo mixdown and the raw audio sources as input and estimate mixing parameters used to create the mixdown by iteratively sampling and *** the sampling step,the inference network predicts a set of mixing parameters,which is sampled and fed to an audio-processing framework to generate audio data for the training *** the training step,the same network used in the sampling step is optimized with the sampled data generated from the sampling *** method is able to explicitly model the mixing process in an interpretable way instead of using a black-box neural network model.A set of objective measures are used for *** experimental results show that this method has better performance than current state-of-the-art methods.

关键词： Audio mixing inversion intelligent audio mixing self-supervised learning audio signal processing deep learning

Role of the hippocampus in contextual modulation of fear extinction

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Neural Regeneration Research 2008年第12期3卷 1386-1389页

作者： Lingzhi Kong xihong wu Liang Li Department of Psychology Peking University Beijing 100871 China Speech and Hearing Research Center Peking University Beijing 100871 China Key Center for Machine Perception and Intelligence Ministry of Education Peking University Beijing 100871 China

Fear extinction is an important form of emotional learning, and affects neural plasticity. Cue fear extinction is a classical form of inhibitory learning that can be used as an exposure-based treatment for phobia, because the long-term extinction memory produced during cue fear extinction can limit the over-expression of fear. The expression of this inhibitory memory partly depends on the context in which the extinction learning occurs. Studies such as transient inhibition, electrophysiology and brain imaging have proved that the hippocampus - an important structure in the limbic system - facilitates memory retrieval by contextual cues. Mediation of the hippocampus-medial prefrontal lobe circuit may be the neurobiological basis of this process. This article has reviewed the role of the hippocampus in the learning and retrieval of fear extinction. Contextual modulation of fear extinction may rely on a neural network consisting of the hippocampus, the medial prefrontal cortex and the amygdala.

关键词： hippocampus fear extinction context-specific exposure-based treatment

维普期刊数据库

A computational model for assessment of speech intelligibility in informational masking

在线全文

学校读者我要写书评

暂无评论

Frontiers of Electrical and Electronic Engineering in China 2012年第1期7卷 107-115页

作者： xihong wu Jing CHEN Speech and Hearing Research Center Key Laboratory of Machine Perception(Ministry of Education)Peking UniversityBeijing 100871China

The existing auditory computational mod- els for evaluating speech intelligibility can only account for energetic masking, and the effect of informational masking is rarely described in these models. This study was aimed to make a computational model considering the mechanism of informational masking. Several psy- choacoustic experiments were conducted to test the ef- fect of informational masking on speech intelligibility by manipulating the number of masking talker, speech rate, and the similarity of F0 contour between target and masker. The results showed that the speech recep- tion threshold for the target increased as the F0 contours of the masker became more similar to that of the tar- get, suggesting that the difficulty in segregating the tar- get harmonics from the masker harmonics may underlie the informational masking effect. Based on these stud- ies, a new auditory computational model was made by inducing the auditory function of harmonic extraction to the traditional model of speech intelligibility index （SII）, named as harmonic extraction （HF） model. The predictions of the HF model are highly consistent with the experimental results.

关键词： auditory computational model speech intelligibility informational masking F0 contour harmonic extraction

Product HMM-based training method for acoustic model with multiple-size units

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Electrical and Electronic Engineering in China 2010年第1期5卷 65-71页

作者： Hao wu xihong wu Huisheng CHI Hearing Research Center Key Laboratory of Machine Perception(Ministry of Education)Peking University Beijing 100871China

Multiple-size units-based acoustic modeling has been proposed for large vocabulary speech recognition system to improve the recognition accuracy with limited training *** introducing a limited number of long-size units into unit set,this modeling scheme can make better acoustic model precision than complete short-size unit modeling without losing model ***,such a multiple-size unit acoustic modeling paradigm does not always bring reliable improvement on recognition performance,since when a large number of long-size units are added in,the amount of training data for short-size units will decrease and result in insufficiently trained *** this paper,a modified Baum-Welch training method is proposed,which uses product hidden Markov models(PHMMs)to couple units with different sizes and enables them to share same portions of training *** validity of proposed method is proved by experiment results.

关键词： multiple-size unit product hidden Markov model(PHMM) product of exports

Software architecture based software deployment reliability estimation considering architectural style

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

High Technology Letters 2012年第3期18卷 281-288页

作者： Su xihong wu Zhibo Liu Hongwei Yang Xiaozong Zuo Decheng School of Computer Science and Technology Harbin Institute of Technology Harbin 150001 P. R. China

Software today often consists of a large number of components offering and requiring services. Such components should be deployed into embedded, pervasive environments, and several deployment architectures are typically possible. These deployment architectures can have significant impacts on system reliability. However, existing reliability estimation approaches are typically limited to certain classes or exclusively concentrate on software reliability, neglecting the influence of hardware resources, software deployment and architectural styles. The selection of an appropriate architectural style has a significant impact on system reliability of the target system. Therefore, we propose a novel software architecture （SA） based reliability estimation model incorporating software deployment and architectural style. On the basis of two architectural styles, we design influence factors and present a new approach to calculate system reliability. Experimental results show that influence factors provide an accurate and simple method of reflecting architectural styles and software deployment on system reliability. It is important for considering the influence of other architectural styles on system reliability in large scale deployment environment.

关键词： software architecture ( SA) software deployment reliability architectural style component

Discriminative training of GMM-HMM acoustic model by RPCL learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Electrical and Electronic Engineering in China 2011年第2期6卷 283-290页

作者： Zaihu PANG Shikui TU Dan SU xihong wu Lei XU Speech and Hearing Research Center Key Laboratory of Machine Perception(Ministry of Education)Peking UniversityBeijing 100871China Department of Computer Science and Engineering The Chinese University of Hong KongHong KongChina

This paper presents a new discriminative approach for training Gaussian mixture models(GMMs)of hidden Markov models(HMMs)based acoustic model in a large vocabulary continuous speech recognition(LVCSR)*** approach is featured by embedding a rival penalized competitive learning(RPCL)mechanism on the level of hidden Markov *** every input,the correct identity state,called winner and obtained by the Viterbi force alignment,is enhanced to describe this input while its most competitive rival is penalized by de-learning,which makes GMMs-based states become more *** the extensive computing burden required by typical discriminative learning methods for one-pass recognition of the training set,the new approach saves computing costs *** show that the proposed method has a good convergence with better performances than the classical maximum likelihood estimation(MLE)based *** with two conventional discriminative methods,the proposed method demonstrates improved generalization ability,especially when the test set is not well matched with the training set.

关键词： discriminative training hidden Markov model rival penalized competitive learning Bayesian Ying-Yang harmony learning large vocabulary continuous speech recognition

SA BASED SOFTWARE DEPLOYMENT RELIABILITY ESTIMATION CONSIDERING COMPONENT DEPENDENCE

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Electronics(China) 2011年第1期28卷 118-125页

作者： Su xihong Liu Hongwei wu Zhibo Yang Xiaozong Zuo Decheng School of Computer Science and Technology Harbin Institute of Technology Harbin 150001 China

Reliability is one of the most critical properties of software *** deployment architecture is the allocation of system software components on host *** Architecture(SA) based software deployment models help to analyze reliability of different *** many approaches for architecture-based reliability estimation exist,little work has incorporated the influence of system deployment and hardware resources into reliability *** are many factors influencing system *** translating the multi-dimension factors into degree matrix of component dependence,we provide the definition of component dependence and propose a method of calculating system reliability of ***,the parameters that influence the optimal deployment may change during system *** existing software deployment architecture may be ill-suited for the given environment,and the system needs to be redeployed to improve *** approximate algorithm,A*_D,to increase system reliability is *** the number of components and host nodes is relative large,experimental results show that this algorithm can obtain better deployment than stochastic and greedy algorithms.

关键词： Software Architecture(SA) Software deployment System reliability Component Redeploy

维普期刊数据库

在线全文

学校读者我要写书评

暂无评论

分块法的模式匹配算法的研究

重庆邮电大学学报（自然科学版） 2014年第4期26卷 551-555页

作者：巫喜红嘉应学院计算机学院广东梅州514015

为提高模式匹配算法性能,介绍经典的模式匹配算法Byoer-Moore和Sunday,分析它们改进后的效率,根据分块法的特点,提出一种新的分块模式匹配(block pattern matching,BPM)算法。BPM算法在预处理阶段先确定模式串的首字符在文本串的位置,... 详细信息

为提高模式匹配算法性能,介绍经典的模式匹配算法Byoer-Moore和Sunday,分析它们改进后的效率,根据分块法的特点,提出一种新的分块模式匹配(block pattern matching,BPM)算法。BPM算法在预处理阶段先确定模式串的首字符在文本串的位置,再确定此字符后长度等于模式串长度的字符是否等于模式串的尾字符,若符合条件,采用单链表存储结构进行存储,在匹配阶段,利用单链表信息进行双向匹配。实验结果表明,BPM算法大大减少了匹配次数和字符比较个数,从而提高匹配效率。

关键词：分块法模式匹配分块模式匹配(BPM)算法 BM算法 Sunday算法

维普期刊数据库