Online comment has become a popular and efficient way for sellers to acquire feedback from customers improve their service quality. However, some key issues need be solved about evaluating improving the hotel quality based on online comments automatically, such as how use less trustworthy comments, discover defects recommend more feasible or economical evaluation indexes comments. To solve above problems, this paper first improves fuzzy comprehensive (FCE) by importing degree it proposes an...

Currently, deep neural networks are deployed on low-power portable devices by first training a full-precision model using powerful hardware, and then deriving corresponding low-precision for efficient inference such systems. However, models directly with coarsely quantized weights is key step towards learning embedded platforms that have limited computing resources, memory capacity, power consumption. Numerous recent publications studied methods networks, but these studies mostly been...

The explosive increasing of the social media data on Web has created and promoted development big mining area welcomed by researchers from both academia industry. sentiment computing news event is a significant component data. It also attracted lot researches, which could support many real-world applications, such as public opinion monitoring for governments recommendation Websites. However, existing methods are mainly based standard emotion thesaurus or supervised methods, not scalable to...

These artificial intelligence technologies have become key tools for solving automated task scheduling as cloud computing applications continue to expand. The paper proposes an automatic scheme large-scale systems based on deep learning and reinforcement learning. addresses the complexity real-time requirements of in such systems. system's parameters are monitored predicted using technology obtain system status information. strategy is then dynamically adjusted state characteristics, a...

Abstract The supply chain network is one of the most important areas focus in majority business circumstances. Blockchain technology a feasible choice for secure information sharing network. Despite fact that maintaining security at all levels blockchain difficult, cryptographic methods are commonly used existing works. Effective management (SCM) offers various benefits to organizations, such as enhanced customer satisfaction, increased operational efficiency, competitive advantage, and cost...

Summary As a well‐known field of big data applications, smart city takes advantage massive analysis to achieve efficient management and sustainable development in the current worldwide urbanization process. An important problem is how discover frequent trajectory sequence pattern cluster trajectory. To solve this problem, paper proposes cloud‐based taxi mining clustering framework for city. Our work mainly includes (1) preprocessing raw Global Positioning System trace by calling Baidu API...

This paper studies how neural network architecture affects the speed of training. We introduce a simple concept called gradient confusion to help formally analyze this. When is high, stochastic gradients produced by different data samples may be negatively correlated, slowing down convergence. But when low, interact harmoniously, and training proceeds quickly. Through theoretical experimental results, we demonstrate confusion, thus efficiency Our results show that, for popular initialization...

Sentiment analysis from microblog platform has received an increasing interest web mining community in recent years. Current sentiment methods are mainly based on the hypothesis that each word expresses only one sentiment. However, human prototyped and fuzzy-confined as declared social psychology, which is conflicting with hypothesis. This of barriers impede computation complex public events microblog. Therefore, how to find a reasonable computational model, combining learning technology...

With the continuous expansion of scale cloud computing applications, artificial intelligence technologies such as Deep Learning and Reinforcement have gradually become key tools to solve automated task scheduling large-scale systems. Aiming at complexity real-time requirement in system, this paper proposes an automatic scheme based on deep learning reinforcement learning. Firstly, technology is used monitor predict parameters system real time obtain status information. Then, combined with...

This paper describes the FAST methodology that enables a single FPGA to accelerate performance of cycle-accurate computer system simulators modeling modem, realistic SoCs, embedded systems and standard desktop/laptop/server systems. The partitions simulator into (i) functional model simulates functionality (ii) predictive predicts other metrics. partitioning is crafted map most parallel work onto hardware-based model, eliminating much complexity difficulty simulating constructs on sequential...

Association link network (ALN) is used to establish associated relations among various resources, aiming at extending the hyperlink World Wide Web an association-rich network, for effectively supporting intelligence activities. Unfortunately, with increase number of challenge incremental building ALN on how perform association weight new coming resources efficiently and exactly. A naive way compare every pair data in existing ALN, thus bearing a O(n <sup...

Cite This Article Z. Xu, "Applications and techniques in cyber intelligence," Computer Systems Science Engineering, vol. 34, no.4, pp. 169–170, 2019.

Depth estimation becomes the key technology to resolve communications of stereo vision. We can get real-time depth map based on hardware, which cannot implement complicated algorithm as software, because there are some restrictions in hardware structure. Eventually, wrong matching will inevitably exist process by such FPGA. In order solve problem a postprocessing function is designed this paper. After cost unique test, both left-right and right-left consistency check solutions implemented,...

We are living in the era of data deluge. Meanwhile, world big includes a rich and complex set cross-media content, including text, images, video, audio, graphics. To date, substantial research efforts have been dedicated to processing mining, yielding good theoretical underpinnings great practical success. However, studies that jointly consider analytics relatively sparse (Zhang, Liu, Deng, Xu, Choo (2018); Zhang, Wei, Wang, Liao (2018)). This gap needs more attention since it will benefit...