As a typical task of multivariate time series, traffic forecasting has always held significant application value across various fields. Most existing models are constrained by fixed input lengths, making it difficult to avoid the ambiguity in spatio-temporal features. We revisited the synergistic effects of Transformer-based models and TCN-MLP models in spatio-temporal data forecasting and proposed the Long-Range Historical Memory-Enhanced transformer (LHMformer). The network consists of a transformer module for processing short-term inputs and a historical feature extraction module for processing long-term inputs. Experiments on six real-world traffic datasets show that the proposed method achieves state-of-the-art results. Follow-up experiments demonstrate that the historical feature extraction module is a key component in solving traffic flow forecasting problems.

Multi-Source cross-lingual transfer learning deals with the transfer of task knowledge from multiple labelled source languages to an unlabeled target language under the language shift. Existing methods typically focus on weighting the predictions produced by language-specific classifiers of different sources that follow a shared encoder. However, all source languages share the same encoder, which is updated by all these languages. The extracted representations inevitably contain different source languages' information, which may disturb the learning of the language-specific classifiers. Additionally, due to the language gap, language-specific classifiers trained with source labels are unable to make accurate predictions for the target language. Both facts impair the model's performance. To address these challenges, we propose a Disentangled and Adaptive Network ~(DA-Net). Firstly, we devise a feedback-guided collaborative disentanglement method that seeks to purify input representations of classifiers, thereby mitigating mutual interference from multiple sources. Secondly, we propose a class-aware parallel adaptation method that aligns class-level distributions for each source-target language pair, thereby alleviating the language pairs' language gap. Experimental results on three different tasks involving 38 languages validate the effectiveness of our approach.

The knowledge distillation-based approaches have recently yielded state-of-the-art (SOTA) results for cross-lingual NER tasks in zero-shot scenarios. These approaches typically employ a teacher network trained with the labelled source (rich-resource) language to infer pseudo-soft labels for the unlabelled target (zero-shot) language, and force a student network to approximate these pseudo labels to achieve knowledge transfer. However, previous works have rarely discussed the issue of pseudo-label noise caused by the source-target language gap, which can mislead the training of the student network and result in negative knowledge transfer. This paper proposes an discrepancy and uncertainty aware Denoising Knowledge Distillation model (DenKD) to tackle this issue. Specifically, DenKD uses a discrepancy-aware denoising representation learning method to optimize the class representations of the target language produced by the teacher network, thus enhancing the quality of pseudo labels and reducing noisy predictions. Further, DenKD employs an uncertainty-aware denoising method to quantify the pseudo-label noise and adjust the focus of the student network on different samples during knowledge distillation, thereby mitigating the noise's adverse effects. We conduct extensive experiments on 28 languages including 4 languages not covered by the pre-trained models, and the results demonstrate the effectiveness of our DenKD.

As cyberattacks on valuable digital assets increase in frequency and sophistication, traditional security gateways are becoming more vulnerable due to their reliance on complex software environments with unpatched vulnerabilities. To address this, we propose a hybrid security gateway that integrates FPGA-based hardware protection to filter out illegitimate traffic and shield the system from external attacks. Our architecture incorporates a novel security protocol based on the Noise Protocol Framework and modern cryptographic algorithms. By modularizing core encryption and filtering tasks and offloading them to independent FPGA components, we reduce the attack surface and enhance overall security. We implemented a prototype and conducted extensive security and usability tests, showing that our system matches the network throughput of traditional gateways while significantly improving resistance to cyber threats, without adding excessive latency.

For named entity recognition (NER) in zero-resource languages, utilizing knowledge distillation methods to transfer language-independent knowledge from the rich-resource source languages to zero-resource languages is an effective means. Typically, these approaches adopt a teacher-student architecture, where the teacher network is trained in the source language, and the student network seeks to learn knowledge from the teacher network and is expected to perform well in the target language. Despite the impressive performance achieved by these methods, we argue that they have two limitations. Firstly, the teacher network fails to effectively learn language-independent knowledge shared across languages due to the differences in the feature distribution between the source and target languages. Secondly, the student network acquires all of its knowledge from the teacher network and ignores the learning of target language-specific knowledge. Undesirably, these limitations would hinder the model's performance in the target language. This paper proposes an unsupervised prototype knowledge distillation network (ProKD) to address these issues. Specifically, ProKD presents a contrastive learning-based prototype alignment method to achieve class feature alignment by adjusting the prototypes' distance from the source and target languages, boosting the teacher network's capacity to acquire language-independent knowledge. In addition, ProKD introduces a prototype self-training method to learn the intrinsic structure of the language by retraining the student network on the target data using samples' distance information from prototypes, thereby enhancing the student network's ability to acquire language-specific knowledge. Extensive experiments on three benchmark cross-lingual NER datasets demonstrate the effectiveness of our approach.

In the face of rising dangers to the internal network due to remote cooperation, VPN gateways are an important tool for organisational network administrators, and the appropriate execution of VPN gateway functions is a vital component in safeguarding the internal network. The VPN gateway confronts security risks from the underlying cryptographic algorithm library, the current operating system, and the central processor as the number of attackers grows and attack methods evolve. In this paper, we propose a pure hardware logic VPN gateway to address security threats from the cryptographic algorithm library, operating system, and CPU by independently implementing the WireGuard protocol’s underlying cryptographic algorithm and building the WireGuard protocol’s hardware logic circuit on the FPGA platform. Actual testing on the NetFPGA-1G-CML platform reveals that the system’s network throughput can reach 35Mbps/s, whereas the network throughput of the software’s WireGuard VPN is 23Mbps/s under the same network settings. Simultaneously, the delay statistics of 300 repetitions of data packet encryption were performed. The encryption latency was less than 20 microseconds when the data packet size was the default MTU.098

Enhancing the interpretability of text classification models can help increase the reliability of these models in real-world applications. Currently, most researchers focus on extracting task-specific words from inputs to improve the interpretability of the model. The competitive approaches exploit the Variational Information Bottleneck (VIB) to improve the performance of word masking at the word embedding layer to obtain task-specific words. However, these approaches ignore the multi-level semantics of the text, which can impair the interpretability of the model, and do not consider the risk of representation overlap caused by the VIB, which can impair the classification performance. In this paper, we propose an enhanced variational word masks approach, named E-VarM, to solve these two issues effectively. The E-VarM combines multi-level semantics from all hidden layers of the model to mask out task-irrelevant words and uses contrastive learning to readjust the distances between representations. Empirical studies on ten benchmark text classification datasets demonstrate that our approach outperforms the SOTA methods in simultaneously improving the interpretability and accuracy of the model.

Automated guided vehicle (AGV) is an important transportation equipment, which is widely used in warehouses and factories. In the scenarios of multi-AGVs application, an efficient AGVs task assignment strategy is beneficial for transportation costs, balance of workload and increasing distribution efficiency. The traditional method usually depends on a powerful scheduling system, which solves the task assignment problem in a regular way. In this paper, we present a decentralized framework of multi-task allocation with attention (MTAA) in deep reinforcement learning, which combines with the methods of task assignment in balance and path planning in cooperation for distribution application. As for task assignment balance, we adopt DNN network to achieve task assignment equilibrium. In multi-AGVs path planning, methods of A3C are embedded in MTAA framework, which are instrumental in improving the stationarity and performance in deep reinforcement learning application. In experiments, we designed two different scenarios under different obstacles to verify the performance of MTAA-A3C and MTAA-DQN methods. The experiments show that the proposed approach has feasibility and effectiveness used in multi-AGVs application.

Communication-enabled Cooperative Adaptive Cruise Control (CACC) is an important technology to improve the safety and traffic capacity of vehicle platoons. However, most existing CACC research considers a conventional communication delay and fixed network topology. When the network is attacked, the stability of the system will be affected. In this paper, a CACC system considering dynamic network topologies and communication delay jitters is proposed. A multi-vehicle look ahead network topology of CACC is designed to investigate the relationship between communication delay and the optimal network topology. Control parameters are obtained by solving Algebraic Riccati equations. Moreover, the relationship between the minimum inter-vehicle spacing and the communication delay is analyzed and derived under different network topologies. Finally, the performance of the proposed CACC system with adaptive network topologies is evaluated by simulations. Simulation results demonstrate that the proposed CACC system outperforms the H infinity synthesis-based controller considering the fixed network topologies. The robustness against communication failure of the proposed system, such as the DeGrading of Service (DGoS) attack, can be significantly improved, which can reduce the minimum safety headway buffer and further support better mobility of vehicles.

Currently, SM9 algorithm has received more and more attention as a new cryptographic product. The SM9 algorithm encryption and decryption principle relies on the mapping relationship on the elliptic curve. Although this mapping relationship improves the security, it will slightly reduce the efficiency. The goal of this article is to improve the efficiency of the SM9 algorithm. Different from the traditional assembly line acceleration method, we decided to start with the basic operation of the algorithm itself. There is a bilinear pairing operation on the elliptic curve, which completes the point to point on the elliptic curve. The calculation complexity directly determines the SM9 algorithm. For this reason, we propose two new bilinear pair processing methods. The former uses the properties of isomorphic mapping to transfer the operations involved in the calculation of bilinear pairs from a large feature domain to a small feature domain, reducing the number of operations on the feature domain. The latter is for special operations in the bilinear pairing process, adding intermediate variables to convert them into low-time-consuming multiplication operations. According to the traditional Miller algorithm, the calculation of bilinear pairs requires 900 multiplication time units. Our solution can reduce this value to 700 and 800 multiplication time units respectively. In addition, the two algorithms have not changed the mapping relationship of the bilinear pair. On the premise of ensuring the correct mapping relationship, the efficiency of the SM9 algorithm is improved.

WireGuard, a novel VPN proposed in 2017, has been widely accepted in the industry, since it is simpler, leaner and more efficient than traditional VPNs. However, WireGuard ignores the risks of key stolen and abused on both desktop and mobile platform. These vulnerabilities might be utilized by attackers to sneak into the protected network through VPN tunnel provided by WireGuard easily without permission. It is especially obvious on mobile platforms, since they are mostly online all day and the attack scenarios are more complicated and changeable. In this paper, we propose an enhanced WireGuard Android version called SeWG based on TEE. We use TEE (Trusted Execution Environment) technology to achieve secure key storage, preventing the misuse of the private key. And we design corresponding dynamic authentication mechanisms for user mode and kernel mode respectively, to prevent malware from sneaking into the secure tunnel. Finally, we implement SeWg scheme on QSEE, the TEE platform developed by Qualcomm. Our experimental results demonstrate that SeWG can work well with high efficiency.

With the development of the wireless body area network, the demand for medical monitoring is getting higher and higher. A series of wearable devices have now emerged as a means of medical monitoring to detect human bioelectrical signals. This article mainly contains three parts, which are wireless body area network, Electrocardiograph(ECG) signal and Electroencephalogram(EEG) signal acquisition, and the design of wearable acquisition system based on wireless body area network. The first part starts with the performance requirements of the wireless body area network, introduces the performance required for medical monitoring, emphasizes the transmission reliability, and focuses on the key issue of wireless body area network research - energy conservation. The second part introduces the non-contact sensor, and expounds the selection of electrode materials. The design of the acquisition circuit and noise cancellation are mainly discussed. The third part puts forward the structural design of the wearable acquisition system based on wireless body area network, and expounds the data transmission system of the system. Finally, it puts forward suggestions and ideas for the design of body-area network-based wearable EEG and ECG acquisition systems.

Cyber-physical systems (CPS) including power systems, transportation, industrial control systems, etc. support both advanced control and communications among system components. Frequent data operations could introduce random failures and malicious attacks or even bring down the whole system. The dependency on a central authority increases the risk of single point of failure. To establish an immutable data provenance scheme for CPS, the authors adopt blockchain and propose a decentralized architecture to assure data integrity. In business-driven CPS, end users are required to share their personal information with multiple third parties. To prevent data leakage and preserve user privacy, the authors isolate and feed different information retrieval requests using tokens specifically generated for each type of request. Providing both traceability of data operations, and unlinkability of end user activities, a robust blockchain-based CPS is prototyped. Evaluation indicates the architecture is capable of assured data provenance validation and user privacy preservation at a low overhead.

The tags are usually tagged by different users in social image sharing websites, which can indicate image semantic information and imply user's preference. Therefore, the tags can contribute to personalized recommendation of social image. However, the present social image tags models only consider single tag, resulting in the relationships among tags are ignored. In this paper, we propose a novel method to create tag tree of social image for personalized recommendation. Firstly, the tag ranking is realized to remove noisy tags. Then, the first layer tags are selected from re-ranked tags lists. To sufficiently express tag's significances, the tag subtrees can be created based on different image categories and combined with first layer tags to create tag tree. Finally, the personalized recommendation of social image is achieved by using tag tree. Experimental results show that our tag tree can effectively express the relationships among tags as well as obtain satisfactory results in personalized recommendation of social image.

Enabled by mobile and wearable technology, personal health data delivers immense and increasing value for healthcare, benefiting both care providers and medical research. The secure and convenient sharing of personal health data is crucial to the improvement of the interaction and collaboration of the healthcare industry. Faced with the potential privacy issues and vulnerabilities existing in current personal health data storage and sharing systems, as well as the concept of self-sovereign data ownership, we propose an innovative user-centric health data sharing solution by utilizing a decentralized and permissioned blockchain to protect privacy using channel formation scheme and enhance the identity management using the membership service supported by the blockchain. A mobile application is deployed to collect health data from personal wearable devices, manual input, and medical devices, and synchronize data to the cloud for data sharing with healthcare providers and health insurance companies. To preserve the integrity of health data, within each record, a proof of integrity and validation is permanently retrievable from cloud database and is anchored to the blockchain network. Moreover, for scalable and performance considerations, we adopt a tree-based data processing and batching method to handle large data sets of personal health data collected and uploaded by the mobile platform.

Although there have been numerous design rationale research projects, few of them has received wider adoption. Because current design rationale systems are not demonstrated the value of the approach in practice due to lack of the design rationale knowledge evaluation. In this paper, we present a new design rationale knowledge evaluation method which is to measure the design rationale knowledge comprehensive value. In this method, we also characterizes the design rationale knowledge value from four perspectives, and the method can provide object metric and selection basis for the design rationale knowledge reuse during the product design process. Partial design process of the bending die is taken as an example to verify the feasibility and validity of the proposed method.

How to achieve a shared understanding among design team members is essential for the collaborative design. Design rationale (DR) explains why and artifact is designed the way it is. An explicit representation of DR can bring many benefits to this objective. This paper argues that the conventional argumentation-based representation of DR has some inherent limitations. Based on this argument, an evolving design rationale (EDR) model is proposed to capture and formalize the DR and its evolving history, in which cognitive elements participating in a design process are organized in an integrated fashion. The model's knowledge structure is detailed and its OWL-based knowledge representation is introduced. The two aspects in which EDR can contribute to the shared understanding are addressed. Finally, a graphical DR modeling system is developed. And the future research based on the proposed model is discussed.

In this paper, the key technologies to the design service collaboration in the cloud manufacturing are explored combining with the cloud manufacturing concept and mode. The design service collaboration mode in the cloud manufacturing environment is proposed. A 5-level 8-view conceptual model and a semantic-based modeling method of the cloud design service are proposed through the analysis of the needs and features of the cloud design service modeling. A service combining method, dynamic generation of the work flow model, is introduced to improve the dynamics and flexibility of the services combining. This can satisfy the needs of the cloud manufacturing mode. Finally, a prototype system is developed to demonstrate the proposed method.

Knowledge is playing a more and more significant role in this knowledge-intensive world, particularly in product-orientated manufacturing in which design knowledge reuse would promote the procedure of new product development. Meanwhile, with increasing applications of Internet and other networked technologies, as well as emerging needs for high-efficiency, energy-saving and service-orientated manufacturing models, Cloud Manufacturing is proposed to meet those needs. In Cloud Manufacturing, all the resources are organized dynamically to provide service-orientated support. However, knowledge is distributed in each braches and subsidiaries of Group Corporation and possesses heterogeneity which would limit the sharing of knowledge across the whole Group Corporation. To deal with those problems, a service-orientated knowledge fusion architecture is presented in this paper. In this architecture, knowledge is packaged as service in the Cloud Manufacturing platform that users could access according to their needs. Base on the knowledge organization model, knowledge fusion could be fulfilled through the ontology mapping and fusion. In the end, conclusion is made about the efficiency of this architecture

Modelica language is been developed to model complex product systems. However, Modelica Models usually are used to simulate the performance not optimize. Computational design methods are very important methods in design process of complex product systems, which is not included in Modelica based tool. In the paper, computational design method integrating design optimization and probabilistic analysis is discussed. The design related elements are analyzed. Then an XML-based method is presented to associate the computational design related elements with Modelica model in a generic way. This method introduces a separation between the model of the system and data related to the design of the product. In order to accomplish full system simulation and optimization involving several types of models and codes, the collaborative process model is presented to represent and execute a computational sequence. Finally, the example about an aircraft landing gear is given.

Engineering product development is becoming increasingly collaborative and knowledge-intensive in today's industry. On one hand, the product development process comprises various tasks that involve extensive knowledge exchange and communication among geographically distributed teams. On the other hand, the product development teams usually reuse engineering knowledge from a vast well of previous design experience when solving new design problems. Hence, there is a strong need to capture, store and reuse the knowledge for product development. In this paper, the different understanding and classifications of engineering knowledge are presented. Two aims of knowledge management for product development are analyzed. Furthermore, the researches of engineering knowledge management are summarized based on critical phases of knowledge management process. Finally, future research directions of knowledge management in engineering are discussed.

Because design is an iterative and evolving process, recording and representing designer's thinking process plays an important role for design change and reuse. After giving a brief review of the previous related research, a design thinking process model is created for representing the design thinking process. Its knowledge structure consisting of design intent, design decision, design rationale and design operation is explained in detail. Based on the proposed model, a knowledge modeling language (Design Mental Modeling Language, DMML) is developed for expressing and reasoning the design thinking process explicitly. Semi-formal structure is used by DMML, which can improve both the expressiveness and computability of this language. A visual modeling system is developed for implementing DMML and an example is presented for illustration.

Shared understanding about both the design content and the design process has significant meaning for collaborative design, which influences the quality of final products. Existing approaches have difficulties in supporting designers to structure and share their design thinking while they are performing real tasks. In this paper we propose Shared Design Thinking Process Model (S-DTPM) to support collaborative design, which is based on co-evolutionary design theory and existing design rationale frameworks. Two principles for S-DTPM have been considered. The elements and representation schema of S-DTPM have been discussed in detail. S-DTPM can help designers to create shared understanding about the design as well as record shared design thinking process for future use such as design review, product maintenance and redesign.

In order to improve the computational efficiency of reliability analysis for MDO, a hierarchical hybrid reliability analysis strategy is proposed. This strategy is composed of two layers, the sequential optimization and reliability assessment (SORA) method is adopted in the upper level to decouple the reliability analysis from MDO, which guarantee that the reliability analysis and deterministic MDO can be executed in a serial way. In the lower layer, the collaborative optimization (CO) strategy is integrated both in deterministic MDO and reliability analysis. Furthermore, based on the traditional theory of most probable point (MPP), the reliability analysis combined with the CO strategy is conducted concurrently. And also, the deterministic MDO is implemented in parallel with the CO strategy. Finally, the proposed strategy is validated with an illustrative example and the results are compared with the existing reliability analysis methods in MDO, showing that the proposed strategy is efficient and competent for evaluating design feasibility in MDO environment with uncertainties.

This paper proposes an integrated system of motor die collaborative manufacturing, and develops a networked public service platform which offers better support for collaborative development and manufacturing of motor die, especially for small and medium sized enterprises, by taking full advantage of the quickly evolving computer network and information technologies. The architecture of the public service platform is presented and the function is designed. The platform is composed of several application service systems: collaborative customer service system, collaborative manufacturing service system, collaborative supply service system and specialized technology service system to promote the closer collaboration among enterprises in an industry chain of motor die. The workflow of the platform is described. Some key technology of the platform is researched. Finally, the implementation of the whole service platform framework with Web technology is introduced.

One way to reduce potentially serious problems caused by modifications to the original design is to find out design thinking process of the original designer. The knowledge structure of design thinking process is represented, which is composed of design intent, design decision, design rationale and design operation. Semi-formal structure is adopted by the presentation, which means both the expressiveness and computability of the representation can be improved. Context in the design thinking process is defined, which can reduce complexity of the design thinking process. Independence between contexts that share an intent allows the designer to propose more assumptions about the design without necessarily affecting previous design thinking process. A graphical modeling system is established for capturing and formalizing the design thinking process.

Assembly related conflicts arising from collaborative product development process severely affect the downstream design and manufacturing activities. A collaborative approach to assembly sequence decision under distributed environment is presented. The product assembly is firstly decomposed into a hierarchical structure tree using an assembly representation based on connectors. The substructures are assigned to different team planners who will generate the sub assembly sequences by various assembly sequence planning systems. By combining these sequences systematically in a bottom-up manner and choosing suitable ones, the assembly sequence of each upper structure including the whole assembly in the structure tree are generated. Using a Web-based system, geographically dispersed team members can perform a 3D assembling visualization of products synchronously and the collaborative assembly sequence decisions are realized in distributed environment. Some prototype systems had been completed and a case was given to illustrate the method.

Assembly planning and assembly sequences combination based on assembly feature interference is proposed and assembly feature interference is introduced as the assembly constraints. Firstly, the assembly features interference is defined and represented. The complete interference is generated by interference expansion of neighboring assembly features' interference. Secondly, the mapping rules between assembly precedence and assembly feature interference is concluded. With the assembly feature interference, assembly planning is performed and assembly sequences combination is also realized by integrating the local precedence of subsequences.