检索结果-内蒙古大学图书馆

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Kun Han Yuxuan Wang DeLiang Wang Department of Computer Science and Engineering & Center for Cognitive and Brain Sciences The Ohio State University Columbus OH 43210-1277 USA

ISBN: (纸本)9781479928941

Reverberation distorts human speech and usually has negative effects on speech intelligibility, especially for hearing-impaired listeners. It also causes performance degradation in automatic speech recognition and speaker identification systems. Therefore, the dereverberation problem must be dealt with in daily listening environments. We propose to use deep neural networks (DNNs) to learn a spectral mapping from the reverberant speech to the anechoic speech. The trained DNN produces the estimated spectral representation of the corresponding anechoic speech. We demonstrate that distortion caused by reverberation is substantially attenuated by the DNN whose outputs can be resynthesized to the dereverebrated speech signal. The proposed approach is simple, and our systematic evaluation shows promising dereverberation results, which are significantly better than those of related systems.

关键词： Speech Speech recognition Reverberation Speech Intelligibility degradation of performance associated system distortion system evaluation negative effects Hearing Disorders

来源：评论

学校读者我要写书评

暂无评论

Hybrid brain-computer interface based on EEG and NIRS modalities

Hybrid brain-computer interface based on EEG and NIRS modali...

引用

International Winter Workshop on brain-computer Interface (BCI)

作者： Min-Ho Lee Siamac Fazli Jan Mehnert Seong-Whan Lee Department of Brain and Cognitive Engineering Korea University Seoul Korea Department of Computer Science Berlin Institute of Technology Berlin Germany

Non-invasive brain-computer interfaces (BCIs) allow users to control external devices by their intentions. Currently, most BCI systems are synchronous, which means, they rely on cues or tasks to which a subject has to react. It would be more useful for users if they could control a device at their own will (i.e., asynchronous BCIs). However, previous asynchronous BCI systems that rely on non-invasive electroencephalogram (EEG) measurements, are not accurate and stable enough for real world applications. Previously, hybrid BCI systems, relying on simultaneous EEG and near-infrared spectroscopy (NIRS) measurements, have been shown to increase the classification performance of synchronous motor imagery (MI) tasks. In this study, we present a first report on an asynchronous multi-modal hybrid BCI, based on simultaneous EEG and near-infrared spectroscopy (NIRS) measurements and propose novel subject-dependent classification strategies for combining both measurements.

关键词： Electroencephalography brain-computer interfaces Accuracy Band-pass filters Digital filters Performance evaluation Visualization

来源：评论

学校读者我要写书评

暂无评论

Neural network based bilingual language model growing for statistical machine translation

Neural network based bilingual language model growing for st...

引用

2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014

作者： Wang, Rui Zhao, Hai Lu, Bao-Liang Utiyama, Masao Sumita, Eiichro Department of Computer Science and Engineering Center for Brain-Like Computing and Machine Intelligence Shanghai Jiao Tong University Shanghai200240 China Multilingual Translation Laboratory MASTAR Project National Institute of Information and Communications Technology 3-5 Hikaridai Keihanna Science City Kyoto619-0289 Japan Key Laboratory Shanghai Education Commission Intelligent Interaction Cognitive Engineering Shanghai Jiao Tong University Shanghai200240 China

ISBN: (纸本)9781937284961

Since larger n-gram Language Model (LM) usually performs better in Statistical Machine Translation (SMT), how to construct efficient large LM is an important topic in SMT. However, most of the existing LM growing methods need an extra monolingual corpus, where additional LM adaption technology is necessary. In this paper, we propose a novel neural network based bilingual LM growing method, only using the bilingual parallel corpus in SMT. The results show that our method can improve both the perplexity score for LM evaluation and BLEU score for SMT, and significantly outperforms the existing LM growing methods without extra corpus. © 2014 Association for Computational Linguistics.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Information geometry meets BCI spatial filtering using divergences

Information geometry meets BCI spatial filtering using diver...

引用

International Winter Workshop on brain-computer Interface (BCI)

作者： Wojciech Samek Klaus-Robert Müller Department of Computer Science Berlin Institute of Technology (TU Berlin) Berlin Germany Department of Brain and Cognitive Engineering Korea University Seoul Republic of Korea

Algorithms using concepts from information geometry have recently become very popular in machine learning and signal processing. These methods not only have a solid mathematical foundation but they also allow to interpret the optimization process and the solution from a geometric perspective. In this paper we apply information geometry to brain-computer Interfacing (BCI). More precisely, we show that the spatial filter computation in BCI can be cast into an information geometric framework based on divergence maximization. This formulation not only allows to integrate many of the recently proposed CSP algorithms in a principled manner, but also enables us to easily develop novel CSP variants with different properties. We evaluate the potentials of our information geometric framework on a data set containing recordings from 80 subjects.

关键词： Covariance matrices Spatial filters Robustness Information geometry Linear programming

来源：评论

学校读者我要写书评

暂无评论

brain-computer interface for smart vehicle: Detection of braking intention during simulated driving

Brain-computer interface for smart vehicle: Detection of bra...

引用

International Winter Workshop on brain-computer Interface (BCI)

作者： Jeong-Woo Kim Il-Hwa Kim Stefan Haufe Seong-Whan Lee Dept. of Brain and Cognitive Engineering Korea Univ. Seoul Korea Dept. of Computer Science Berlin Institute of Technology Berlin Germany

Neuro-driving simulation framework was proposed in this article for studying neural correlates of braking intention in diversified driving situations. In addition, the possibility was investigated that these neural correlates can be used to detect a participant's braking intention prior to the behavioral response. Electroencephalographic (EEG) and electromyographic (EMG) signals were measured from fifteen participants during they were exposed to several kinds of traffic situations in a neuro-driving simulation framework. After that, the novel characteristic feature was extracted from the measured signals to categorize according to whether the driver intended to brake or not. This proposed novel feature consists of readiness potential (RP), event-related desynchronization (ERD) and event-related potential (ERP) as used in a previous study. The prediction performance of braking intention based on the proposed feature combination exhibited superior prediction performance than simple ERP feature used in a previous study.

关键词： Vehicles Feature extraction Electroencephalography Accidents brain modeling Electromyography computers

来源：评论

学校读者我要写书评

暂无评论

Information-theoretic objective functions for lifelong learning

Information-theoretic objective functions for lifelong learn...

引用

2013 AAAI Spring Symposium

作者： Zhang, Byoung-Tak School of Computer Science and Engineering Graduate Programs in Cognitive Science and Brain Science Seoul National University Seoul 151-742 Korea Republic of

ISBN: (纸本)9781577356028

Conventional paradigms of machine learning assume all the training data are available when learning starts. However, in lifelong learning, the examples are observed sequentially as learning unfolds, and the learner should continually explore the world and reorganize and refine the internal model or knowledge of the world. This leads to a fundamental challenge: How to balance long-term and short-term goals and how to trade-off between information gain and model complexity? These questions boil down to "what objective functions can best guide a lifelong learning agent?" Here we develop a sequential Bayesian framework for lifelong learning, build a taxonomy of lifelong-learning paradigms, and examine information-theoretic objective functions for each paradigm, with an emphasis on active learning. The objective functions can provide theoretical criteria for designing algorithms and determining effective strategies for selective sampling, representation discovery, knowledge transfer, and continual update over a lifetime of experience. © 2013, Association for the Advancement of artificial intelligence.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

Data and information granule rules retrieval: Differences of activation in parietal cortex

Data and information granule rules retrieval: Differences of...

引用

2014 International Conference on brain Informatics and Health, BIH 2014

作者： Zhao, Wei Li, Hongyu Gu, Gue Wang, Xiuzhen Zhou, Guohui Cui, Jiaxin Gu, Weiquan College of Computer Science and Information Engineering Harbin Normal University Harbin China Software College Northeast Agricultural University Harbin China State Key Laboratory of Cognitive Neuroscience and Learning IDG/McGovern Institute for Brain Research Beijing Normal University Beijing China Center for Collaboration and Innovation in Brain and Learning Sciences Beijing Normal University Beijing China Harbin Normal University Harbin China

ISBN: (纸本)9783319098906

Efficient encoding of Roman rules is based on the neural bases of mathematical cognitive abilities. The present imaging studies have shown that information granule representing a form of Roman rules is associated with arithmetical domain-sensitive parietal cortex, indicating a switch from the data to the information granule retrieval of memory rules. So far, however, little is known about the developing neural substrate for the establishment of rules from data to information granule. The aim of the present fMRI study is to investigate whether and how mathematical intelligence might be enhanced from data to information granule of Roman arithmetic rules in the parietal cortex. Concerning the same rules, the paired t-test analysis indicated that different activation in the bilateral parietal lobule associated with different retrieval levels. In conclusion, the present study yielded some evidence that a successful model for knowledge-building of rules is accompanied by modifications of brain activation patterns. © 2014 Springer International Publishing.

关键词： Chemical activation

来源：评论

学校读者我要写书评

暂无评论

Modeling crowd motions for abnormal activity detection

Modeling crowd motions for abnormal activity detection

引用

IEEE Conference on Advanced Video and Signal Based Surveillance (AVSS)

作者： Dong-Gyu Lee Heung-Il Suk Seong-Whan Lee Department of Computer Science and Engineering Korea University Anam-dong Seongbuk-Ku Seoul Korea Department of Brain and Cognitive Engineering Korea University Anam-dong Seongbuk-Ku Seoul Korea

In this paper; we propose a novel crowd behavior representation method to detect abnormal behaviors in videos. An adaptive optical flow filtering method is proposed to utilize low-level optical flow informations. Furthermore, a simple framework is developed to detect and to localize abnormal crowd behavior using adaptive optical flow filtering result. The proposed method is more robust than other modeling methods in representing different behaviors. In this model, a normal behavior is presented by the general value. Some outliers in the temporal domain or spatial domain are presented by a higher value. Spatio-temporal cuboids are extracted from the filtering result to present the likelihood of anomaly in the frame. Experimental evaluations are performed on two public datasets with comparison to the provisos abnormal behavior detection methods in the literature. Experimental results show that the proposed methods outperform previous abnormal behavior detection techniques in the literature.

关键词： Feature extraction Adaptive optics Optical filters Optical imaging Force Optical distortion Dynamics

来源：评论

学校读者我要写书评

暂无评论

Instantaneous monitoring of heart beat dynamics during anesthesia and sedation

引用

Journal of Computational Surgery 2014年第1期1卷 1-18页

作者： Gaetano Valenza Oluwaseun Akeju Kara J Pavone Luca Citi Katharine E Hartnack Aaron Sampson Patrick L Purdon Emery N Brown Riccardo Barbieri Department of Anesthesia Critical Care and Pain Medicine Massachusetts General Hospital Boston USA Harvard Medical School Boston USA Department of Information Engineering University of Pisa Pisa Italy School of Computer Science and Electronic Engineering University of Essex Colchester UK Department of Brain and Cognitive Sciences Massachusetts Institute of Technology Cambridge USA

Anesthesia-induced altered arousal depends on drugs having their effect in specific brain regions. These effects are also reflected in autonomic nervous system (ANS) outflow dynamics. To this extent, instantaneous monitoring of ANS outflow, based on neurophysiological and computational modeling, may provide a more accurate assessment of the action of anesthetic agents on the cardiovascular system. This will aid anesthesia care providers in maintaining homeostatic equilibrium and help to minimize drug administration while maintaining antinociceptive effects. In previous studies, we established a point process paradigm for analyzing heartbeat dynamics and have successfully applied these methods to a wide range of cardiovascular data and protocols. We recently devised a novel instantaneous nonlinear assessment of ANS outflow, also suitable and effective for real-time monitoring of the fast hemodynamic and autonomic effects during induction and emergence from anesthesia. Our goal is to demonstrate that our framework is suitable for instantaneous monitoring of the ANS response during administration of a broad range of anesthetic drugs. Specifically, we compare the hemodynamic and autonomic effects in study participants undergoing propofol (PROP) and dexmedetomidine (DMED) administration. Our methods provide an instantaneous characterization of autonomic state at different stages of sedation and anesthesia by tracking autonomic dynamics at very high time-resolution. Our results suggest that refined methods for analyzing linear and nonlinear heartbeat dynamics during administration of specific anesthetic drugs are able to overcome nonstationary limitations as well as reducing inter-subject variability, thus providing a potential real-time monitoring approach for patients receiving anesthesia.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Estimating Multiple Evoked Emotions from Videos 35

Estimating Multiple Evoked Emotions from Videos

引用

35th Annual Meeting of the cognitive science Society - Cooperative Minds: Social Interaction and Group Dynamics, CogSci 2013

作者： Choe, Wonhee Chun, Hyo-Sun Noh, Junhyug Lee, Seong-Deok Zhang, Byoung-Tak Cognitive Science Program Seoul National University Seoul151-744 Korea Republic of R&D Center Samsung Electronics Suwon 443-742 Korea Republic of School of Computer Science and Engineering Seoul National University Seoul151-744 Korea Republic of Yongin449-712 Korea Republic of Computer Science and Engineering Cognitive Science and Brain Science Programs Seoul National University Seoul151-744 Korea Republic of

ISBN: (纸本)9780976831891

Video-sharing websites have begun to provide easy access to user-generated video content. How do we find what we want to view among the huge video database? When people search for a video, they may want to know whether the video evokes a certain emotional sensation. The evoked emotion is one of the important factors we consider when we select a video. One of the key concepts of evoked emotions from videos: the evoked emotions are different for each scene and for each viewer. Considering these differences, we obtained human-evoked emotions from 33 videos. We used these emotions to estimate the multiple emotions evoked by each scene of the videos. Using a computational model of emotion estimation based on mid-level visual features, we found that, in individual videos, the same scene evoked multiple emotions. Our results show that a video evoked different emotions from different people. A computational model might deliver probabilistic multiple-evoked emotions from video analyses. © CogSci *** rights reserved.

关键词： Video recording

来源：评论

学校读者我要写书评

暂无评论

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：

请选择收藏分类：

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

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：