Current Issue : July - September Volume : 2019 Issue Number : 3 Articles : 5 Articles
We propose a novel end-to-end approach, namely, the semantic-containing double-level embedding Bi-LSTM model (SCDE-Bi-\nLSTM), to solve the three key problems of Q&A matching in the Chinese medical field. In the similarity calculation of the Q&A core\nmodule, we propose a text similarity calculation method that contains semantic information, to solve the problem that previous Q&A\nmethods do not incorporate the deep information of a sentence into the similarity calculations. For the sentence vector representation\nmodule, we present a double-level embedding sentence representation method to reduce the error caused by Chinese medical word\nsegmentation. In addition, due to the problem of the attention mechanism tending to cause backward deviation of the features, we\npropose an improved algorithm based on Bi-LSTM in the feature extraction stage. The Q&A framework proposed in this paper not\nonly retains important timing features but also loses low-frequency features and noise. Additionally, it is applicable to different\ndomains. To verify the framework, extensive Chinese medical Q&A corpora are created. We run several state-of-the-art Q&A\nmethods as contrastive experiments on the medical corpora and the current popular insuranceQA dataset under different performance\nmeasures. The experimental results on the medical corpora show that our framework significantly outperforms several\nstrong baselines and achieves an improvement of top-1 accuracy of up to 14%, reaching 79.15%....
Motivated by the effects of deregulation over power quality and the subsequent need of\nnew types of measurements, this paper assesses different implementations of an estimate for the\nspectral kurtosis, considered as a low-level harmonic detection. Performance of a processor-based\nsystem is compared with a field programmable gate array (FPGA)-based solution, in order to\nevaluate the accuracy of this processing function for implementation in autonomous measurement\nequipment. The fourth-order spectrum, with applications in different fields, needs advanced\ndigital signal processing, making it necessary to compare implementation alternatives. In order to\nobtain reproducible results, the implementations have been developed using common design and\nprogramming tools. Several characteristics of the implementations are compared, showing that the\nincreasing complexity and reduced cost of the current FPGA models make the implementation of\ncomplex mathematical functions feasible. We show that FPGAs improve the processing capability\nof the best processor using an operating frequency 33 times lower. This fact strongly supports its\nimplementation in hand-held instruments....
This paper presents design details adopting open embedded systems (OES) as real-time\ncontrollers in industrial distributed control systems. OES minimize development cost and enhance\nportability while addressing widely known shortcomings of their proprietary counterparts.\nThese shortcomings include the black box method of distribution which hinders integration to\nmore complex systems. However, OES are highly dependent on the compatibility of each software\ncomponents and essential benchmarking is required to ensure that the system can satisfy hard\nreal-time constraints. To address these issues and the notion that OES will find broader distributed\ncontrol applications, we provide detailed procedures in realizing OES based on an open source\nreal-time operating system on various low-cost open embedded platforms. Their performance\nwas evaluated and compared in terms of periodicity and schedulability, task synchronization, and\ninterrupt response time, which are crucial metrics to determine stability and reliability of real-time\ncontrollers. Practical implementations, including the modernization of a multi-axis industrial robot\ncontroller, are described clearly to serve as a comprehensive reference on the integration of OES in\nindustrial distributed control systems....
The development of the Internet of Things has made embedded devices widely used. Embedded devices are often used to process\nsensitive data, making them the target of attackers. ARM TrustZone technology is used to protect embedded device data from\ncompromised operating systems and applications. But as the value of the data stored in embedded devices increases, more and\nmore effective physical attacks have emerged.However, TrustZone cannot resist physical attacks.We propose SoftME, an approach\nthat utilizes the on-chip memory space to provide a trusted execution environment for sensitive applications. We protect the\nconfidentiality and integrity of the data stored on the off-chip memory. In addition, we design task scheduling in the encryption\nprocess. We implement a prototype system of our approach on the development board supporting TrustZone and evaluate the\noverhead of our approach.The experimental results show that our approach improves the security of the system, and there is no\nsignificant increase in system overhead....
One of the most widely used models for specifying functional requirements is a use case\nmodel. The viewpoint of the use case model that views a system as a black box focuses on descriptions\nof external interactions between the system and related environments. However, for embedded\nsystems that do not disclose most implementation logics outside the system, black box-based use case\nmodels may experience the drawback that considerable information that must be defined for system\ndevelopments is omitted. To solve this shortcoming, several studies have been proposed on the use\nof kind of white box technique in which the dynamic behaviors of embedded systems are defined\nfirst using a state diagram and the results are reflected in the requirement specifications. However,\nwhite box-based modeling has not been widely adopted by developers due to tasks that require a lot\nof time in the requirement analysis phase in the initial phase of the software development life cycle.\nThis study proposes a gray box-based requirement specification method as a trade-off between two\ncontradictory elements (the amount of information required to develop an embedded system and the\ncost of the effort required during the requirement analysis phase) in terms of the two approaches, the\nblack and the white box-based models. The proposed method suggests that an appropriate depth\nlevel of embedded system modeling is required to define the requirements. This study also proposes\na mechanism that automatically generates an application programming interface for each component\nbased on the created model. The proposed method was applied to the development of a camera\nsensor controller in a mobile phone, and the case results proved the feasibility of the method through\ndiscussion of the application results....
Loading....