In service engineering it is important to estimate when and what a worker did, because they include crucial evidences to improve service quality and working environments. For Service Operation Estimation (SOE), acoust...
详细信息
ISBN:
(纸本)9781479988518
In service engineering it is important to estimate when and what a worker did, because they include crucial evidences to improve service quality and working environments. For Service Operation Estimation (SOE), acoustic information is one of useful and key modalities;particularly environmental or background sounds include effective cues. This paper focuses on two aspects: (1) extracting powerful and robust acoustic features by using stacked-denoising-autoencoder and bag-of-feature techniques, and (2) investigating a multi-modal SOE scheme by combining the audio features and the other sensor data as well as non-sensor information. We conducted evaluation experiments using multi-modal data recorded in a restaurant. We improved SOE performance in comparison to conventional acoustic features, and effectiveness of our multi-modal SOE scheme is also clarified.
The purpose of quantum state tomography（QST） is to obtain a complete quantum state by reconstructing a density matrix from experimental data and therefore gives researchers a powerful tool to analyze complex synthet...
详细信息
The purpose of quantum state tomography（QST） is to obtain a complete quantum state by reconstructing a density matrix from experimental data and therefore gives researchers a powerful tool to analyze complex synthetic quantum *** the past decades,many methods have been developed to improve the efficiency of quantum state tomography（QST）,such as old methods: compress sensing on almost pure state and low-rank systems,new method: using restricted Boltzmann machine on systems of quantum ***,they are both limited to some certain *** this paper,we introduce an universal method using deep neural networks based on stacked denoising autoencoder,which probably can help finish QST on all types of systems with a much smaller number of measurements and therefore is very efficient.
State Assessment and Fault Prediction mechanism of distribution terminal is the premise of ensuring safe and reliable operation of power grid. However, the sample size of fault rate data of distribution terminal is us...
详细信息
ISBN:
(纸本)9781728149318
State Assessment and Fault Prediction mechanism of distribution terminal is the premise of ensuring safe and reliable operation of power grid. However, the sample size of fault rate data of distribution terminal is usually small and the data is missing seriously, so it is difficult to accurately evaluate and analyze the fault rate data. A reliability assessment and prediction model based on stacked denoising autoencoder and hierarchical bayesian is proposed in this paper. Firstly, the quantile method is used to detect abnormal values of distribution terminals. Then, features of fault data is extracted by stacked denoising autoencoder(SDAE), and the Hierarchical Bayesian(HB) model is used to evaluate the failure rate data of distribution terminals. Finally, the evaluation method is compared the traditional prior distribution and the second-order linear regression. The results show that the proposed method improves the accuracy of fault assessment by 5.37% and 10.09%, the accuracy of prediction by 10.23% and 19.94%. The proposed method has higher reliability, prediction efficiency and evaluation accuracy.
stackedautoencoder(SAE) is hard to achieve satisfactory performance,when input data are complex and ***,the identification performance of recurrent neural network(RNN) may decrease rapidly under noisy *** order to de...
详细信息
stackedautoencoder(SAE) is hard to achieve satisfactory performance,when input data are complex and ***,the identification performance of recurrent neural network(RNN) may decrease rapidly under noisy *** order to deal with these problems,a novel hybrid deep neural network(DNN) based on stacked denoising autoencoder(SDAE) and gated recurrent unit neural network(GRUNN) is ***,the structure of the presented hybrid DNN is *** hybrid DNN contains a SDAE,a GRUNN,and a softmax ***,the training algorithm based on action discovery(AD) is proposed to train the presented hybrid *** experimental studies indicate the presented hybrid DNN processes strong anti-noise ability and adaptability to time-varying signals.
暂无评论