Search

Most researchers focus on designing accurate crowd counting models with heavy parameters and computations but ignore the resource burden during the model deployment. A real-world scenario demands an efficient counting model with low-latency and high-performance. Knowledge distillation provides an elegant way to transfer knowledge from a complicated teacher model to a compact student model while maintaining accuracy. However, the student model receives the wrong guidance with the supervision of the teacher model due to the inaccurate information understood by the teacher in some cases. In this paper, we propose a dual-knowledge distillation (DKD) framework, which aims to reduce the side effects of the teacher model and transfer hierarchical knowledge to obtain a more efficient counting model. First, the student model is initialized with global information transferred by the teacher model via adaptive perspectives. Then, the self-knowledge distillation forces the student model to learn the knowledge by itself, based on intermediate feature maps and target map. Specifically, the optimal transport distance is utilized to measure the difference of feature maps between the teacher and the student to perform the distribution alignment of the counting area. Extensive experiments are conducted on four challenging datasets, demonstrating the superiority of DKD. When there are only approximately 6% of the parameters and computations from the original models, the student model achieves a faster and more accurate counting performance as the teacher model even surpasses it. [ABSTRACT FROM AUTHOR]

With the development of intelligent applications, more and more intelligent applications are computation intensive, data intensive and delay sensitive. Compared with traditional cloud computing, edge computing can reduce communication delay by offloading computing tasks to edge cloud. Furthermore, with the complexity of computing scenarios in edge cloud, deep learning based on computation offloading scheme has attracted wide attention. However, all the learning-based offloading scheme does not consider the where and how to run the offloading scheme itself. Thus, in this paper, we consider the problem of running the learning-based computation offloading scheme for the first time and propose the learning for smart edge architecture. Then, we give the computation offloading optimization problem of mobile devices under multi-user and multi edge cloud scenarios. Furthermore, we propose cognitive learning-based computation offloading (CLCO) scheme for this problem. Finally, experimental results show that compared with other offloading schemes, the CLCO scheme has lower task duration and energy consumption. [ABSTRACT FROM AUTHOR]

With the development of technologies such as the Internet of Things and artificial intelligence, mobile applications are becoming more and more intelligent. Compared to traditional applications realized by mobile cloud computing technology, these novel applications have a higher requirement for a task offloading scheme. However, the traditional task offloading schemes are hard pressed to meet latency and personalization requirements of these new applications. For intelligent application, how to realize personalized and fine-grained task offloading is still a challenging problem. Therefore, we propose a scheme called intelligent task offloading (iTaskOffloading) for a cloud-edge collaborative system, which can provide personalized task offloading. To be specific, we first propose the architecture of iTaskOffloading which includes the local device layer, edge cloud layer, remote cloud layer, and cognitive engine. Then we analyze the method of iTaskOffloading, which contains coarse-grained computing and fine-grained computing. Finally, we build a testbed to evaluate the proposed iTaskOffloading scheme using a typical intelligent application of emotion detection. The experimental results show that compared to the traditional cloud computing scheme, iTaskOffloading has less task duration. [ABSTRACT FROM AUTHOR]

The development of mobile communication technology, hardware, distributed computing, and artificial intelligence (AI) technology has promoted the application of edge computing in the field of heterogeneous IoT in order to overcome the defects of the traditional cloud computing model in the era of big data. In this article, we first propose a new AI-enabled smart edge with heterogeneous IoT architecture that combines edge computing, caching, and communication. Then we propose the Smart-Edge-CoCaCo algorithm. To minimize total delay and confirm the computation offloading decision, Smart-Edge-CoCaCo uses joint optimization of the wireless communication model, the collaborative filter caching model in edge cloud, and the computation offloading model. Finally, we built an emotion interaction testbed to perform computational delay experiments in real environments. The experiment results show that the computation delay of the Smart-Edge-CoCaCo algorithm is lower than that of the traditional cloud computing model with the increase of computing task data and the number of concurrent users. [ABSTRACT FROM AUTHOR]

Highlights • Propose a deep weight fusion method for emotion recognition. • Conduct the cross-modal noise modeling for the audio and video data. • Use the deep belief network for multi-modal emotion features. Abstract The multi-modal emotion recognition lacks the explicit mapping relation between emotion state and audio and image features, so extracting the effective emotion information from the audio/visual data is always a challenging issue. In addition, the modeling of noise and data redundancy is not solved well, so that the emotion recognition model is often confronted with the problem of low efficiency. The deep neural network (DNN) performs excellently in the aspects of feature extraction and highly non-linear feature fusion, and the cross-modal noise modeling has great potential in solving the data pollution and data redundancy. Inspired by these, our paper proposes a deep weighted fusion method for audio-visual emotion recognition. Firstly, we conduct the cross-modal noise modeling for the audio and video data, which eliminates most of the data pollution in the audio channel and the data redundancy in visual channel. The noise modeling is implemented by the voice activity detection(VAD), and the data redundancy in the visual data is solved through aligning the speech area both in audio and visual data. Then, we extract the audio emotion features and visual expression features via two feature extractors. The audio emotion feature extractor, audio-net, is a 2D CNN, which accepting the image-based Mel-spectrograms as input data. On the other hand, the facial expression feature extractor, visual-net, is a 3D CNN to which facial expression image sequence is feeded. To train the two convolutional neural networks on the small data set efficiently, we adopt the strategy of transfer learning. Next, we employ the deep belief network(DBN) for highly non-linear fusion of multi-modal emotion features. We train the feature extractors and the fusion network synchronously. And finally the emotion classification is obtained by the support vector machine using the output of the fusion network. With consideration of cross-modal feature fusion, denoising and redundancy removing, our fusion method show excellent performance on the selected data set. [ABSTRACT FROM AUTHOR]

With the development of intelligent applications (e.g., self-driving, real-time emotion recognition), there are higher requirements for cloud intelligence. However, cloud intelligence depends on the multi-modal data collected by user equipments (UEs). Due to the limited capacity of network bandwidth, offloading all data generated from the UEs to the remote cloud is impractical. Thus, in this article, we consider the challenging issue of achieving a certain level of cloud intelligence while reducing network traffic. In order to solve this problem, we design a traffic control algorithm based on label-less learning on the edge cloud, which is dubbed LLTC. By use of the limited computing and storage resources at the edge cloud, LLTC evaluates the value of data that will be offloaded. Specifically, we first give a statement of the problem and the system architecture. Then we design the LLTC algorithm in detail. Finally, we set up the system testbed. Experimental results show that the proposed LLTC can guarantee the required cloud intelligence while minimizing the amount of data transmission. [ABSTRACT FROM AUTHOR]

This paper introduces basic concept of mood fatigue detection and existing solutions as well as some typical solutions, such as mobile sensing and cloud computing technology. In the first place, we sum up main challenges of mood fatigue detection and the direction of future study. Then one type of system implementation is put forward, such system consists of: 1) Wearable devices used to detect the users' mood fatigue; 2) Clouds data center; 3) Application and service manager. We take overall consideration of many factors like the user's physiological information, external environment conditions and user behavior, evaluate accurately through big data analytic technology the users' state of mood fatigue and to what extent shall one keeps vigilant as well as what measures shall be adopted to improve the users' working performance and alert the users to keep themselves away from the danger. Finally a real system is established in this paper, it is composed of the smart clothing, cloud platform and mobile terminal application. [ABSTRACT FROM AUTHOR]

As mobile devices are equipped with more memory and computational capability, a novel peer-to-peer communication model for mobile cloud computing is proposed to interconnect nearby mobile devices through various short range radio communication technologies to form mobile cloudlets, where every mobile device works as either a computational service provider or a client of a service requester. Though this kind of computation offloading benefits compute-intensive applications, the corresponding service models and analytics tools are remaining open issues. In this paper we categorize computation offloading into three modes: remote cloud service mode, connected ad hoc cloudlet service mode, and opportunistic ad hoc cloudlet service mode. We also conduct a detailed analytic study for the proposed three modes of computation offloading at ad hoc cloudlet. [ABSTRACT FROM PUBLISHER]

In this paper, we present a mathematical model of an infectious disease according to the characteristics of the COVID-19 pandemic. The proposed enhanced model, which will be referred to as the SEIR (Susceptible–Exposed–Infectious–Recovered) model with population migration, is inspired by the role that asymptomatic infected individuals, as well as population movements can play a crucial role in spreading the virus. In the model, the infected and the basic reproduction numbers are compared under the influence of intervention policies. The experimental simulation results show the impact of social distancing and migration-in rates on reducing the total number of infections and the basic reproductions. And then, the importance of controlling the number of migration-in people and the policy of restricting residents' movements in preventing the spread of COVID-19 pandemic are verified. • Establish a mathematical model of an infectious disease based on the characteristics of COVID-19. • Calculate the curve of the number of infections based on the Euler numerical method. • Analyze and compare the effects of different anti-epidemic policies on reducing the number of infections. [ABSTRACT FROM AUTHOR]

With the continuous expansion of data scale, data condensation technology has emerged as a means to reduce costs related to storage, time, and energy consumption. Data condensation can generate a synthesized dataset of reduced size, enabling the training of models that exhibit high performance comparable to the original dataset. Nevertheless, data condensation has also exposed privacy issues. Although many approaches have been proposed to preserve privacy for data condensation, the privacy protection for data condensation has not been well explored. Furthermore, to the best of our knowledge, none of the existing approaches propose dynamic parameters-based differential privacy dataset condensation considering unnecessary noise introduced by the fixed privacy parameter strategy. Most approaches typically inject constant noise with the fixed variance into gradients across all layers using predefined privacy parameters, which can significantly impact model accuracy. In this paper, we investigate alternative approaches for data condensation with differential privacy (DP) that aim to ensure DP while minimizing the noise added to gradients and improving the model accuracy. First, we develop a dynamic threshold method to reduce the noise added to gradients in the later stages of training by using a clipping threshold that decreases with training rounds. Second, noise injection in our method is not arbitrary as in conventional approaches; instead, it is based on the maximum size of the gradient after clipping. This approach ensures that only minimal noise increments are introduced, thereby mitigating accuracy loss and parameter instability that may arise from excessive noise injection. Finally, our privacy analysis confirms that the proposed method provides a rigorous privacy guarantee. Extensive evaluations on different datasets demonstrate that our approach can improve accuracy compared to existing DP data condensation techniques while adhering to the same privacy budget and applying a specified clipping threshold. [ABSTRACT FROM AUTHOR]

With rapid development of unmanned aerial vehicle (UAV) technology, application of UAVs for task offloading has received increasing interest in academia. However, real-time interaction between one UAV and the mobile edge computing node is required for processing the tasks of mobile end users, which significantly increases the system overhead and is unable to meet the demands of large-scale artificial intelligence (AI)-based applications. To tackle this problem, in this article, we propose a new architecture for UAV clustering to enable efficient multi-modal multi-task offloading. With the proposed architecture, the computing, caching, and communication resources are collaboratively optimized using AI-based decision making. This not only increases the efficiency of UAV clusters, but also provides insight into the fusion of computation and communication. [ABSTRACT FROM AUTHOR]

Recent advances in wireless networking and big data technologies, such as 5G networks, medical big data analytics, and the Internet of Things, along with recent developments in wearable computing and artificial intelligence, are enabling the development and implementation of innovative diabetes monitoring systems and applications. Due to the life-long and systematic harm suffered by diabetes patients, it is critical to design effective methods for the diagnosis and treatment of diabetes. Based on our comprehensive investigation, this article classifies those methods into Diabetes 1.0 and Diabetes 2.0, which exhibit deficiencies in terms of networking and intelligence. Thus, our goal is to design a sustainable, cost-effective, and intelligent diabetes diagnosis solution with personalized treatment. In this article, we first propose the 5G-Smart Diabetes system, which combines the state-of-the-art technologies such as wearable 2.0, machine learning, and big data to generate comprehensive sensing and analysis for patients suffering from diabetes. Then we present the data sharing mechanism and personalized data analysis model for 5G-Smart Diabetes. Finally, we build a 5G-Smart Diabetes testbed that includes smart clothing, smartphone, and big data clouds. The experimental results show that our system can effectively provide personalized diagnosis and treatment suggestions to patients. [ABSTRACT FROM AUTHOR]

According to a report by the World Health Organization, diseases caused by an unhealthy lifestyle represent the leading cause of death all over the world. Therefore, it is crucial to monitor and avoid users' unhealthy behaviors. Existing health monitoring approaches still face many challenges of limited intelligence due to insufficient healthcare data. Therefore, this article proposes a smart personal health advisor (SPHA) for comprehensive and intelligent health monitoring and guidance. The SPHA monitors both physiological and psychological states of the user. The SPHAScore model is proposed to evaluate the overall health status of the user. Finally, a testbed for verification of feasibility and applicability of the proposed system was developed. The experimental and simulation results have shown that the proposed approach is efficient for proper user state monitoring. [ABSTRACT FROM AUTHOR]

With the emergence of mobile edge computing (MEC), the edge cloud with certain computing power is deployed closer to the mobile device, which can well solve the computing and delay requirements of the mobile device. In 5G ultra-dense heterogeneous networks, where the macro base station (MBS) and multiple dense small base stations (SBS) are deployed in the region, the offloading decision faces multiple choices. In order to solve the problem of computing offloading and resource allocation in 5G ultra-dense heterogeneous networks, we propose a collaborative optimization strategy based on multi-agent deep reinforcement learning (MADRL). At each time, the mobile device only needs to make the optimal offloading decision according to its own historical offloading decision, the allocated bandwidth and computing resources at the past time, as well as the service response delay and energy consumption at the past time, without knowing other user information and dynamic network environment information. Simulation results show that the proposed collaborative optimization strategy is better than the other three baseline schemes in terms of service response delay and energy consumption performance. [Display omitted] • A joint optimization strategy of task offloading and resource allocation for 5G ultra-dense networks is proposed. • The offloading decision of mobile devices is made based on MAPPO. • The heterogeneous edge clouds realize the allocation of bandwidth and computing resources based on convex optimization. [ABSTRACT FROM AUTHOR]