Children with Autism Spectrum Disorder (ASD) often struggle with social communication and feel anxious in interactive situations. The Picture Exchange Communication System (PECS) is commonly used to enhance basic communication skills in children with ASD, but it falls short in reducing social anxiety during therapist interactions and in keeping children engaged. This paper proposes the use of virtual character technology alongside PECS training to address these issues. By integrating a virtual avatar, children's communication skills and ability to express needs can be gradually improved. This approach also reduces anxiety and enhances the interactivity and attractiveness of the training. After conducting a T-test, it was found that PECS assisted by a virtual avatar significantly improves children's focus on activities and enhances their behavioral responsiveness. To address the problem of poor accuracy of gaze estimation in unconstrained environments, this study further developed a visual feature-based gaze estimation algorithm, the three-channel gaze network (TCG-Net). It utilizes binocular images to refine the gaze direction and infer the primary focus from facial images. Our focus was on enhancing gaze tracking accuracy in natural environments, crucial for evaluating and improving Joint Attention (JA) in children during interactive processes.TCG-Net achieved an angular error of 4.0 on the MPIIGaze dataset, 5.0 on the EyeDiap dataset, and 6.8 on the RT-Gene dataset, confirming the effectiveness of our approach in improving gaze accuracy and the quality of social interactions.

In the context of the increasingly diversified blockchain technology, interoperability among heterogeneous blockchains has become key to further advancing this field. Existing cross-chain technologies, while facilitating data and asset exchange between different blockchains to some extent, have exposed issues such as insufficient security, low efficiency, and inconsistent standards. Consequently, these issues give rise to significant obstacles in terms of both scalability and seamless communication among blockchains within a multi-chain framework. To address this, this paper proposes an efficient method for cross-chain interaction in a multi-chain environment. Building upon the traditional sidechain model, this method employs smart contracts and hash time-locked contracts (HTLCs) to design a cross-chain interaction scheme. This approach decentralizes the execution of locking, verifying, and unlocking stages in cross-chain transactions, effectively avoiding centralization risks associated with third-party entities in the process. It also greatly enhances the efficiency of fund transfers between the main chain and sidechains, while ensuring the security of cross-chain transactions to some extent. Additionally, this paper innovatively proposes a cross-chain data interaction strategy. Through smart contracts on the main chain, data from sidechains can be uploaded, verified, and stored on the main chain, achieving convenient and efficient cross-chain data sharing. The contribution of this paper is the development of a decentralized protocol that coordinates the execution of cross-chain interactions without the need to trust external parties, thereby reducing the risk of centralization and enhancing security. Experimental results validate the effectiveness of our solution in increasing transaction security and efficiency, with significant improvements over existing models. Our experiments emphasize the system's ability to handle a variety of transaction scenarios with improved throughput and reduced latency, highlighting the practical applicability and scalability of our approach.

Consumer electronics technology significantly shapes the landscape of virtual worlds, with devices like virtual reality headsets, augmented reality glasses, smartphones, and haptic feedback devices providing users with immersive experiences. However, the rapid evolution of the metaverse raises concerns about privacy and security as digitization expands into every aspect of the real world, resulting in a substantial increase in data volume. To address these concerns, blockchain technology emerges as a potential solution due to its decentralized nature and tamper-proof capabilities. Blockchain can play a vital role in safeguarding personal information and virtual identities within the metaverse. Nevertheless, existing consensus protocols in large systems may incur high network and storage costs. To address this issue, this paper proposes a Raft consensus protocol for adaptive data maintenance in the metaverse based on Cauchy Reed-Solomon (CRS) codes. This protocol reduces the data storage requirements of nodes by leveraging erasure code techniques and optimizes blockchain processing capacity by adding coded data blocks to functioning servers. To demonstrate the advantages of the scheme, we construct a key-value pair based on ACRS-Raft and evaluate it. The experimental results demonstrate that ACRS-Raft saves about 70% of storage space and improves write throughput by 260% over the Raft protocol.

The Java virtual machine (JVM) is the cornerstone of the Java platforms. A JVM's exception handling implementation interrupts, when the objective application encounters an exception (or an error), the normal execution of the application and performs specific handling tasks. However, little research has been done in systematically validating JVMs' exception handling implementations—test programs or even applications need to be carefully designed for throwing/catching exceptions at runtime; a JVM's exception handling implementation is also complicated, making it challenging to design tests for testing all of its functionalities. Inspired by the recent success of fuzz testing of compilers and JVM implementations, we introduce ehcbuilder , the first technique for fuzzing JVMs' exception handling implementations. The key idea is to construct exception handling chains , each of which abstracts a program's execution into a sequence of exception throwings, catchings, and/or handlings. A classfile seed can then be mutated into test programs with diverse exception handling chains, enabling (1) exceptions to be continuously thrown and caught at runtime, and (2) JVMs' exception handling implementations to be much more thoroughly tested. We have implemented ehcbuilder and evaluated ehcbuilder on popular JVM implementations including OpenJDK's HotSpot, Eclipse's OpenJ9, Azul's Zulu, and Oracle's GraalVM. Our results show that ehcbuilder can generate programs with very intricate exception handling chains and reveal differences among JVMs' exception handling implementations: Up to thousands of lines of source code in HotSpot's exception handling implementation are covered more than the original benchmarks; during 39 K iterations, ehcbuilder generates exception handling chains of different lengths, revealing 258 runtime differences. We classify the differences into four categories, and reveal a fast throw issue confirmed by HotSpot developers and another initcause issue confirmed by the OpenJ9 community.

Purpose Customization, a marketing strategy through providing personalized products, might be a new solution to motivate consumer feedbacks in electronic commerce (e-commerce) websites. Taking the dual-value of customization (emotional involvement and uniqueness expression) as the theoretical basis, this study aims to investigate the impact of customization on consumer's word-of-mouth (WOM) behaviors and contents by motivating: (1) more, (2) faster, (3) positive at first and then negative, (4) longer and (5) more helpful WOMs. Design/methodology/approach A field study was conducted with multi-sourced data: customer order data from a Chinese retailer and WOM data from Amazon.com. The two datasets were matched to filter out 463 online reviews among 6,892 customers who placed customized orders. Heckman's two-stage model, logistic regression, Ordinary least squares regression, Tobit regression, analysis of covariance and Lind–Mehlum U Test were used in the data analysis. Findings This study has found that (1) customization level motivates WOM behaviors including WOM posting and WOM speed, (2) an inverted U -shaped relationship exists between customization level and consumer rating and (3) customization level has a significantly positive impact on WOM helpfulness but not on WOM length. Originality/value This study advances theoretical development in the area of WOM motivators by proposing a new product-centric approach, customization, to stimulate voluntary WOMs. Empirical field research that analyzes consumer's real responses to customization is in scarcity. The dual-value of customized products is proposed as the underlying mechanism to explain the impact of customization level on consumer's WOM behaviors/contents. An interesting inverted U -shaped relationship is found between customization level and customer rating. This research provides nuanced practical guidance for websites, companies and consumers.

随着攻击技术的不断演进,防御的难度也与日俱增.为了及时、有效地识别和阻断攻击的实施,学术界与工业界已提出众多基于攻击检测的防御技术.现有的攻击检测方法主要着眼于攻击事件,通过识别攻击特征或者定位异常活动来发现攻击,分别具有泛化性和攻击导向性不足的局限性,容易被攻击者精心构造的攻击变种绕过,造成漏报和误报.然而,根据观察发现:尽管攻击及其变种可能采用众多不同的攻击机制来绕过一些防御措施,以实现同一攻击目的,但是由于攻击目的不变,这些攻击对系统的影响依然具有相似性,因此,所造成的系统影响并不会随攻击手段的大量增多而随之产生对应的增长.针对这一特点,提出了基于攻击指标依赖模型的攻击检测方法,以更有效地应对攻击变种.所提出的指标依赖模型着眼于漏洞利用后对系统的影响而非变化多样的攻击行为,因此具有更强的泛化能力.基于模型指导,进一步采用多层次监控技术,以迅速捕获定位攻击迹,最终实现对目标攻击与变种的精确检测,有效降低攻击检测的误报率.在DARPA透明计算项目以及典型APT攻击组成的测试集上,与现有的基于攻击事件分析的检测方法进行实验对比,结果表明:在预设场景下,所提出的方法可以根据可接受的性能损耗实现99.30％的检出率.

With the continued increasing availability of spatio-temporal data from GPS-equipped devices, online map services, and a variety of location-based social media, trajectory similarity search has become a fundamental operation in location-based data analytics. It is of great importance to enable real-time search of trajectories that satisfy users’ personalized requirements. For the purpose, we study the Diversified Continuous Trajectory Similarity Search (DCTSS) problem. The DCTSS problem aims to process a large number of Continuous Location Set (CLS) queries over a stream of trajectory data while taking result diversity of each query into consideration. To answer the DCTSS problem, we develop a Diversity-Aware Trajectory Publish/Subscribe (DAT-PS) framework, which takes both trajectory data streams and CLS queries as input and performs query-trajectory matching between CLS queries in the query collection and trajectories over the trajectory data stream. Our experimental results on two real-life datasets show that our proposed DAT-PS framework is capable of demonstrating substantial superiority regarding both efficiency and scalability compared against baselines.

Spatiotemporal rules are ubiquitous nowadays. For example, a user may need to pay for a phone call or the usage of the Internet services. The payment may depend on many factors such as the duration of the call, the locations of the caller and callee, etc. Thus, the rules are spatiotemporal in nature. Given a set of such spatiotemporal rules and the behavior of a user, the problem of spatiotemporal rule matching (STRM) is to find the activated rules for the behavior. In this paper, we consider user behaviors as special instances of spatiotemporal data and formalize the STRM problem. We show how to represent common rules as spatiotemporal rules. Then, we show how to process multiple rules in advance to generate non-conflicting rules and organize them into tree-based index structures. After that, we propose novel algorithms, taking advantage of the tree-based index structures, to solve the STRM problem. Extensive experiments show that our proposed algorithms can outperform the baseline methods by a large margin, and thus are applicable in large scale real-life scenarios.

With the rapid development of information technology, the devel-opment of blockchain technology has also been deeply impacted. When performing block verification in the blockchain network, if all transactions are verified on the chain, this will cause the accumulation of data on the chain, resulting in data storage problems. At the same time, the security of data is also challenged, which will put enormous pressure on the block, resulting in extremely low communication efficiency of the block. The traditional blockchain system uses the Merkle Tree method to store data. While verifying the integrity and correctness of the data, the amount of proof is large, and it is impossible to verify the data in batches. A large amount of data proof will greatly impact the verification efficiency, which will cause end-to-end communication delays and seriously affect the blockchain system's stability, efficiency, and security. In order to solve this problem, this paper proposes to replace the Merkle tree with polynomial commitments, which take advantage of the properties of polynomials to reduce the proof size and communication consumption. By realizing the ingenious use of aggregated proof and smart contracts, the verification efficiency of blocks is improved, and the pressure of node communication is reduced.

Industry 5.0 is seen as the next stage of the industrial revolution. However, the centralized management of the vast amount of data in Industry 5.0 poses significant security challenges. Because blockchain can provide a stable decentralized network for data in Industry 5.0, blockchain technology becomes an ideal solution for the security of data in Industry 5.0. However, with the explosion of data in Industry 5.0, the rapidly growing storage requirements are putting unbearable pressure on nodes and users, making it especially important to address the scalability of the blockchain. While using light nodes can alleviate storage pressure on nodes, it can also lead to data availability attacks. To address the above issues, this paper proposes a blockchain-based trusted data storage mechanism for Industry 5.0. The mechanism combines sharding and two-layer Merkle tree structures, and employs random low-density parity codes to code the tree. It enables light nodes to verify the authenticity of data and solve data availability attacks while improving the throughput of the blockchain. In addition, by using erasure codes to construct low-storage blockchain nodes, the network load of the nodes is effectively reduced, which solves the scalability issues of the blockchain.

As more and more mobile devices and IoT terminals are connected to the Internet, a huge amount of spatio-temporal data is generated. In order to cope with the pressure of storing and computing massive amounts of data locally and to share data to fully utilize its value, data owners usually store data in cloud servers. However, the data owner will lose physical control of big spatio-temporal data while the big spatio-temporal data is stored on the cloud, and it will be at the risk of being tampered with and deleted. Therefore, this paper designs a blockchain-based secure storage mechanism, BSMD, which adopts an on-chain and off-chain cooperative storage model to alleviate the shortage of blockchain storage capacity. The updatable subvector commitment designed in this paper is used to construct the on-chain and off-chain secure authentication protocol which ensures the consistency of on-chain and off-chain data and makes the protocol have the capability of batch processing. Finally, the correctness and security of the proposed protocol are proved, and the performance of the protocol is analyzed.

Distributed storage can store data in multiple devices or servers to improve data security.However, in today's explosive growth of network data, traditional distributed storage scheme is faced with some severe challenges such as insufficient performance, data tampering, and data lose.A distributed storage scheme based on blockchain has been proposed to improve security and efficiency of traditional distributed storage.Under this scheme, the following improvements have been made in this paper.This paper first analyzes the problems faced by distributed storage.Then proposed to build a new distributed storage blockchain scheme with sharding blockchain.The proposed scheme realizes the partitioning of the network and nodes by means of blockchain sharding technology, which can improve the efficiency of data verification between nodes.In addition, this paper uses polynomial commitment to construct a new verifiable secret share scheme called PolyVSS.This new scheme is one of the foundations for building our improved distributed storage blockchain scheme.Compared with the previous scheme, our new scheme does not require a trusted third party and has some new features such as homomorphic and batch opening.The security of VSS can be further improved.Experimental comparisons show that the proposed scheme significantly reduces storage and communication costs.

Traffic signs are one of the most important information that guide cars to travel, and the detection of traffic signs is an important component of autonomous driving and intelligent transportation systems. Constructing a traffic sign dataset with many samples and sufficient attribute categories will promote the development of traffic sign detection research. In this paper, we propose a new Chinese traffic sign detection benchmark, which adds more than 4,000 real traffic scene images and corresponding detailed annotations based on our CCTSDB 2017, and replaces many original easily-detected images with difficult samples to adapt to the complex and changing detection environment. Due to the increase of the number of difficult samples, the new benchmark can improve the robustness of the detection network to some extent compared to the old version. At the same time, we create new dedicated test sets and categorize them according to three aspects: category meanings, sign sizes, and weather conditions. Finally, we present a comprehensive evaluation of nine classic traffic sign detection algorithms on the new benchmark. Our proposed benchmark can help determine the future research direction of the algorithm and develop a more precise traffic sign detection algorithm with higher robustness and real-time performance.

Data augmentation has been proved effective in training deep models. Existing data augmentation methods tackle fine-grained problem by blending image pairs and fusing corresponding labels according to the statistics of mixed pixels, which produces additional noise harmful to the performance of networks. Motivated by this, we present a simple yet effective cross ensemble knowledge distillation (CEKD) model for fine-grained feature learning. We innovatively propose a cross distillation module to provide additional supervision to alleviate the noise problem, and propose a collaborative ensemble module to overcome the target conflict problem. The proposed model can be trained in an end-to-end manner, and only requires image-level label supervision. Extensive experiments on widely used fine-grained benchmarks demonstrate the effectiveness of our proposed model. Specifically, with the backbone of ResNet-101, CEKD obtains the accuracy of 89.59%, 95.96% and 94.56% in three datasets respectively, outperforming state-of-the-art API-Net by 0.99%, 1.06% and 1.16%.

The short-term heavy rainfalls, thunderstorm gales, hail, squall lines, tornadoes, thunderstorms and other disastrous weather caused by deep convective clouds greatly threaten social and economic activities and the safety of people's lives and property. Therefore, it is of great value to study the recognition methods of deep convective clouds in the field of the weather forecast. Since deep convective clouds are characterized by a short life cycle, small spatial scale and complex structure, it is difficult to accurately monitor and identify deep convective clouds by traditional ground monitoring methods. In this paper, the semantic segmentation network SCNET based on attention mechanism was proposed and a deep learning network for the recognition of deep convective clouds was established with infrared and brightness temperature channels of FY4A stationary meteorological satellites as input features. The results showed that SCNET has a better recognition effect than meteorological and machine learning methods, such as single-band threshold method, SVM, NN, UNET and RESNET, and can effectively improve the recognition accuracy of deep convective clouds

With the rapid development of industrialization, the environmental pollution issue is becoming increasingly serious, especially the air pollution problem. As the core of the prevention and control of air pollution, air pollution prediction plays a very significant role in human survival and development. Therefore, it is highly essential to develop an accurate air pollution prediction model for mass rallies (e.g., playground and bazaar). Recent studies have suggested that multiple air contaminants, e.g., PM 2.5 and PM 10 , which belong to a kind of aerosol, can carry the Covid-19 virus and spread it rapidly through the atmosphere, and this dramatically increases the risk of Covid-19 infection, particularly in the crowded and enclosed environment. Nevertheless, most existing air pollution prediction methods, which rely on large amounts of historical data for modeling and assume that the crowd flows relatively slow, are difficult to apply well to predict air pollution in mass rallies. To solve the aforementioned problem and better assist the decision-makers in managing environmental risk to human beings, in this article, we come up with a novel air pollution prediction model for mass rallies. More specifically, we first propose a temporally weighting matrix to differentiate the significance of training samples in the time domain. Then, we construct a temporal support vector regressor (TSVR), which puts more emphasis on the adjacent samples by considering the fact that the crowd usually flows promptly and disorderly in mass rallies. Finally, based on the extended TSVR, we develop a multitask TSVR (MTSVR) that simultaneously considers the related tasks. Since different air contaminants are correlated with each other, all the tasks can benefit by sharing information. The results of comparison experiments demonstrate that our presented MTSVR outperforms state-of-the-art single-task learners, multitask learners, and air pollution predictors when applied for air pollution prediction in mass rallies. Particularly, when under the six-task condition, the error values of the prediction of PM 2.5 , PM 10 , and O 3 obtained by our proposed method are relatively lower, outperforming the most advanced method tested by 15.2%, 6.1%, and 4.3%, and the precision values of the predicted values outperform the advanced method tested by 28.3%, 25.1%, and 24.8%.

Parkinson's disease (PD) is known as a rampant neurodegenerative disorder, which has afflicted approximately 10 million people throughout the world. Surface Electromyography (sEMG) signal trials received from the upper extremities, such as the arm and wrist, would be an efficient way to assess neuromuscular function in the detection of PD. This paper mainly aimed to utilize pre-trained deep transfer learning (DTL) structures and conventional machine learning (ML) models as an automated approach to diagnose PD from sEMG signals. Primarily, we stacked the extracted features from three deep pre-trained architectures, including AlexNet, VGG-f, and CaffeNet, to generate the discriminative feature vectors. Although the number of stacked features from all the three deep structures was large, the proper features is effective in overcoming the challenge of over-fitting as well as increasing the robustness to added noise with different levels. Subsequently, we proposed a novel soft combination of subset feature selection methods, including receiver operating characteristic (ROC), entropy, and the signal-to-noise (SNR) procedures, in order to reduce the size of the extracted features. Finally, we utilized the support vector machine (SVM) with radial basis function (RBF) kernel for identifying PD disorder. The experimental results in different analysis frameworks illustrated that the hybrid deep transfer learning-based approach to PD classification could lead to hitting rates higher than 99%. Moreover, it can be of a competitive performance with the state-of-the-art SVM-based pattern even though the suggested model needs minimal processing in feature construction of sEMG signals to PD detection.

Due to the dim light in the tunnel and the characteristics of natural electromagnetic shielding, drivers are prone to accidents in the tunnel and are unable to inform the navigation system in time. Therefore, it is still a bottleneck in the field of intelligent transportation how to obtain real-time vehicle position information and vehicle state information when the vehicle is running at high speeds in the tunnel. In this paper, a new technology is proposed to achieve accurate real-time positioning in the tunnel scene by combining the downlink time difference of arrival (DL-TDOA) with UWB technology. The DL-TDOA technology is based on Ultra Wide Band (UWB) data transmission technology, which can effectively reduce the interference of other electromagnetic waves and reduce the data transmission time. By calculating the transmission time of the wireless electromagnetic wave between the vehicle and the fixed base station, the technology can determine the real-time position of the vehicle and greatly reduce the time loss of data in transmission. DL-TDOA based on UWB technology has high precision, while DL-TDOA based on UWB technology has many advantages, such as high precision, strong anti-jamming ability, low power consumption and a high transmission rate which are suitable for accurate positioning and navigation in tunnel scenarios. In the final tunnel experiment, several tests were carried out at speeds of 30km/h, 60km/h and 80km/h respectively. By comparing the coordinate position after the conversion with the satellite coordinate, the real-time kinematics (RTK), it was concluded that the position error of vehicles in the tunnel is less than 1m, and the real-time positioning of vehicles in the tunnel is realized.

Multi-carrier energy systems have received wide attentions due their flexibility and sustainable characteristics. Although these systems show significant efficiency in providing and consuming energy, the performance of the whole system can be degraded owing to uncertainties arising from different sources. This paper presents a stochastic decentralized model for considering the uncertainties of a system including different types of thermal and electrical private loads using a multi-agent framework. In other words, agents have private ownership and seek for social welfare as well as optimized personal profits. In the proposed model, the gradient projection method is used to implement a fully-decentralize energy trading model. Also, various stochastic scenarios of solar irradiance, prices, and loads are considered using the fast-forward selection algorithm to take into account the uncertainties. Then, to assess the proposed stochastic multi-agent model, "AnyLogic" is used for conducting the simulation studies. The numerical results show that the clearing price is directly affected by the renewable agent without any supervisory control. Moreover, the total operating cost of the considered multi-carrier energy system decreases by similar to 7 % considering these uncertainties compared with a deterministic one. However, social welfare declines due to the intrinsically beneficial behavior in private cooperation.

As the typical peer-to-peer distributed networks, blockchain systems require each node to copy a complete transaction database, so as to ensure new transactions can by verified independently. In a blockchain system (e.g., bitcoin system), the node does not rely on any central organization, and every node keeps an entire copy of the transaction database. However, this feature determines that the size of blockchain transaction database is growing rapidly. Therefore, with the continuous system operations, the node memory also needs to be expanded to support the system running. Especially in the big data era, the increasing network traffic will lead to faster transaction growth rate. This paper analyzes blockchain transaction databases and proposes a storage optimization scheme. The proposed scheme divides blockchain transaction database into cold zone and hot zone using expiration recognition method based on Least Recently Used (LRU) algorithm. It can achieve storage optimization by moving unspent transaction outputs outside the in-memory transaction databases. We present the theoretical analysis on the optimization method to validate the effectiveness. Extensive experiments show our proposed method outperforms the current mechanism for the blockchain transaction databases.