期刊：
IEEE ROBOTICS AND AUTOMATION LETTERS,2024年9(3):2646-2653 ISSN：2377-3766
通讯作者：
Lu, ZY
作者机构：
[Zhao, Zhou] Cent China Normal Univ, Sch Comp Sci, Wuhan 430079, Peoples R China.;[Lu, Zhenyu; He, Wenhao] Univ West England, Fac Environm & Technol, Bristol BS16 1QY, England.;[Lu, Zhenyu; He, Wenhao] Univ West England, Bristol Robot Lab, Bristol BS16 1QY, England.
通讯机构：
[Lu, ZY ] U;Univ West England, Fac Environm & Technol, Bristol BS16 1QY, England.;Univ West England, Bristol Robot Lab, Bristol BS16 1QY, England.
关键词：
Grasping;Robots;Robot sensing systems;Tactile sensors;Deep learning;Exoskeletons;Sensors;deep learning in grasping and manipulation;learning from experience
摘要：
To minimize irrelevant and redundant information in tactile data and harness the dexterity of human hands. In this paper, we introduce a novel binary classification network with normalized differential convolution (NDConv) layers. Our method leverages the recent progress in visual-based tactile sensing to significantly improve the accuracy of grasp stability prediction. First, we collect a dataset from human demonstration by grasping 15 different daily objects. Then, we rethink pixel correlation and design a novel NDConv layer to fully utilize spatio-temporal information. Finally, the classification network not only achieves a real-time temporal sequence prediction but also obtains an average classification accuracy of 92.97%. The experimental results show that the network can hold a high classification accuracy even when facing unseen objects.
摘要：
Pseudo-label (PL) learning-based methods usually regard class confidence above a certain threshold for unlabeled samples as PLs, which may result in PLs still containing wrong labels. In this letter, we propose a prototype-based PL refinement (PPLR) for semi-supervised hyperspectral image (HSI) classification. The proposed PPLR filters wrong labels from PLs using class prototypes, which can improve the discrimination of the network. First, PPLR uses multihead attentions (MHAs) to extract the spectral-spatial features, and designs an adaptive threshold that can be dynamically adjusted to generate high-confidence PLs. Then, PPLR constructs class prototypes for different categories using labeled sample features and unlabeled sample features with refined PLs to improve the quality of PLs by filtering wrong labels. Finally, PPLR further assigns reliable weights (RWs) to these PLs in calculating their supervised loss, and introduces a center loss (CL) to improve the discrimination of features. When ten labeled samples per category are utilized for training, PPLR achieves the overall accuracies of 82.11%, 86.70%, and 92.50% on the Indian Pines (IP), Houston2013, and Salinas datasets, respectively.
作者机构：
[Zhang, Yang; Zhang, Y] Shanghai Polytech Univ, Sch Econ & Management, Shanghai 201209, Peoples R China.;[Wang, Xuechun] Hubei Univ Technol, Sch Elect & Elect Engn, Wuhan 430068, Peoples R China.;[Wen, Jinghao] Cent China Normal Univ, Sch Comp Sci, Wuhan 430079, Peoples R China.;[Zhu, Xianxun] Shanghai Univ, Sch Commun & Informat Engn, Shanghai 200444, Peoples R China.
通讯机构：
[Zhang, Y ] S;Shanghai Polytech Univ, Sch Econ & Management, Shanghai 201209, Peoples R China.
关键词：
Human presence sensing;Machine learning;Non-contact;Wireless perception
摘要：
<jats:title>Abstract</jats:title><jats:p>In the swiftly evolving landscape of Internet of Things (IoT) technology, the demand for adaptive non-contact sensing has seen a considerable surge. Traditional human perception technologies, such as vision-based approaches, often grapple with problems including lack of sensor versatility and sub-optimal accuracy. To address these issues, this paper introduces a novel, non-contact method for human presence perception, relying on WiFi. This innovative approach involves a sequential process, beginning with the pre-processing of collected Channel State Information (CSI), followed by feature extraction, and finally, classification. By establishing signal models that correspond to varying states, this method enables the accurate perception and recognition of human presence. Remarkably, this technique exhibits a high level of precision, with sensing accuracy reaching up to 99<jats:inline-formula><jats:alternatives><jats:tex-math>$$\%$$</jats:tex-math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML">
<mml:mo>%</mml:mo>
</mml:math></jats:alternatives></jats:inline-formula>. The potential applications of this approach are extensive, proving to be particularly beneficial in contexts such as smart homes and healthcare, amongst various other everyday scenarios. This underscores the significant role this novel method could play in enhancing the sophistication and effectiveness of human presence detection and recognition systems in the IoT era.</jats:p>
通讯机构：
[Liang, P ] W;Wuhan Univ, Sch Comp Sci, Wuhan, Peoples R China.;Hubei Luojia Lab, Wuhan, Peoples R China.
关键词：
Mining Architectural Information;Software Repositories;Architecting Activity;Software Development;Systematic Mapping Study
摘要：
Mining Software Repositories (MSR) has become an essential activity in software development. Mining architectural information (e.g., architectural models) to support architecting activities, such as architecture understanding, has received significant attention in recent years. However, there is a lack of clarity on what literature on mining architectural information is available. Consequently, this may create difficulty for practitioners to understand and adopt the state-of-the-art research results, such as what approaches should be adopted to mine what architectural information in order to support architecting activities. It also hinders researchers from being aware of the challenges and remedies for the identified research gaps. We aim to identify, analyze, and synthesize the literature on mining architectural information in software repositories in terms of architectural information and sources mined, architecting activities supported, approaches and tools used, and challenges faced. A Systematic Mapping Study (SMS) has been conducted on the literature published between January 2006 and December 2022. Of the 104 primary studies finally selected, 7 categories of architectural information have been mined, among which architectural description is the most mined architectural information; 11 categories of sources have been leveraged for mining architectural information, among which version control system (e.g., GitHub) is the most popular source; 11 architecting activities can be supported by the mined architectural information, among which architecture understanding is the most supported activity; 95 approaches and 56 tools were proposed and employed in mining architectural information; and 4 types of challenges in mining architectural information were identified. This SMS provides researchers with promising future directions and help practitioners be aware of what approaches and tools can be used to mine what architectural information from what sources to support various architecting activities.
作者机构：
[Xie, Zhen] Henan Univ Sci & Technol, Business Sch, Luoyang 471000, Peoples R China.;[Feng, Hua; Bu, Wei] Beijing Jiaotong Univ, Sch Econ & Management, Beijing 100044, Peoples R China.;[Wang, Yan] Cent China Normal Univ, Sch Comp, Wuhan 430079, Peoples R China.
通讯机构：
[Xie, Z ] H;Henan Univ Sci & Technol, Business Sch, Luoyang 471000, Peoples R China.
关键词：
Internet of Things;Smart manufacturing;Blockchain;Cloud-edge collaborative computing;High-tech manufacturing
摘要：
This study proposes an IoT-based smart manufacturing method, which aims to improve the integration efficiency of the industrial chain and innovation chain in high-tech manufacturing. By introducing blockchain technology and cloud-edge collaborative computing optimization, our method shows significant advantages in data integrity, system response time, computing resource utilization, cost-effectiveness, data security, system interoperability, supply chain collaborative efficiency, product quality, customer satisfaction and data processing efficiency. The simulation results show that the proposed method is superior to the traditional method in all key indicators, especially in improving data integrity (98%), reducing system response time (250ms), improving computing resource utilization (CPU utilization 85%) and significantly reducing costs. By optimizing the data processing process and algorithm, our method can process and analyze big data more efficiently, thereby improving the overall performance and response speed of the system, and providing a solid theoretical and practical foundation for the intelligent transformation of high-tech manufacturing.
摘要：
Entities and relations extraction are the key tasks in the construction of biomedical knowledge graph, which play an important role in the biomedical artificial intelligence. However, extraction of entities and relations from biomedical texts is challenging because of the overlapping triples problem. The previous approaches typically divided the task into two separate sub-tasks. However, these methods failed to address the error propagation problem. Recent methods have been proposed to perform both sub-tasks simultaneously. Nonetheless, most current methods still encounter issues related to imbalanced interactions and independent features. In this paper, we propose a novel method based on feature partition encoding and relative positional embedding to joint extract biomedical entity and relation triples simultaneously. Compared to previous work, our method shows exceptional accurate in extracting entities and relations, while efficiently tackling the challenge of overlapping triples in biomedical texts. Our work has two contributions. Firstly, our method divides the features into task-specific and shared parts through entity, relation and sharing partitions at the encoding stage. And the encoded features will be aggregated according to the subsequent tasks. Secondly, we introduce a relative positional embedding method to capture the relative distance information between token pairs. In this way, our method can effectively deal with the sub-tasks interactions problem and improve entities and relations extraction. The experimental results show that our method improves the F1 scores of relations extraction by 3.2%, 2.1%, 3.4%, and 2.8% on four biomedical datasets, respectively.
摘要：
Visual Dialog aims to answer an appropriate response based on a multi-round dialog history and a given image. Existing methods either focus on semantic interaction, or implicitly capture coarse-grained structural interaction (e.g., pronoun co-references). The fine-grained and explicit structural interaction feature for dialog history is seldom explored, resulting in insufficient feature learning and difficulty in capturing precise context. To address these issues, we propose a structure-aware dual-level graph interactive network (SDGIN) that integrates verb-specific semantic roles and co-reference resolution to explicitly capture context structural features for discriminative and generative tasks in visual dialog. Specifically, we create a novel structural interaction graph that injects syntactic knowledge priors into dialog by introducing semantic role labeling that imply which words are sentence stems. Furthermore, considering the single perspective limitation of previous algorithms, we design a dual-perspective mechanism that learns fine-grained token-level context structure features and coarse-grained utterance-level interactions in parallel. It possess an elegant view to explore precise context interactions, realizing the mutual complementation and enhancement of different granularity features. Experimental results show the superiority of our proposed approach. Compared to other task-specific models, our SDGIN outperforms previous models and achieves a significant improvement on the benchmark dataset VisDial v1.0.
摘要：
Inspired by the concept of divide-and-conquer, existing multi-task/ multi-population constraint evolutionary algorithms (CMOEAs) have often employed an auxiliary population that disregards all constraints in order to simplify the problem. However, when dealing with complex Constraint Pareto Fronts (CPF), many existing approaches encounter difficulties in maintaining diversity and avoiding local optima. To address the above issue, the Three-role-community based CMOEA (TRC) which focuses on roles within the population is introduced to eliminate the burden of knowledge transfer between multi-task or multi-population CMOEAs. TRC establishes three essential roles: the feasible group, tasked with identifying CPFs; the exploration group, dedicated to discovering the unconstrained Pareto Front (UPF); and the diversity group, responsible for preserving population diversity. By dynamically adjusting the allocation of individuals to these roles, TRC effectively navigates the evolving problem landscape. Moreover, a flexible and straightforward quota allocation strategy for offspring size is designed in TRC. Rigorously tested on MW and DASCMOP test suites, TRC's performance is either better than or at least comparable to some state-of-the-art algorithms.
摘要：
With the rapid development of IoT technology, smart homes have emerged. At the same time, data security and privacy protection are also of great concern. However, the traditional centralized authentication scheme has defects such as single point of failure, poor scalability, center dependence, vulnerability to attacks, etc., and is not suitable for the distributed and dynamically changing smart home environment. Thus, many researchers have proposed decentralized authentication schemes based on blockchain technology. Although many characteristics of blockchain technology such as decentralization, non -tampering, and solving single point of failure have good application scenarios in authentication, the mature integration of the two applications has to be further explored. For example, the introduction of blockchain also brings security issues; the balance between security and performance in most blockchain-based authentication schemes remains to be investigated; and resource -constrained IoT devices tend to perform a large number of intensive computations, which is clearly inappropriate. Consequently, this paper introduces fog computing in blockchain-based authentication schemes, proposes a network architecture in which cloud and fog computing work together, and investigates the security and performance issues of authentication schemes under this architecture. Finally, formal and informal security analysis show that our scheme has multiple security properties, and our scheme has better performance than existing solutions.
期刊：
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS,2024年PP:1-12 ISSN：2168-2194
作者机构：
[Xueli Pan; Frank van Harmelen] Department of Computer Science, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands;Hubei Provincial Key Laboratory of Artificial Intelligence and Smart Learning, Central China Normal University, Wuhan, China;National Engineering Research Center for E-Learning, Central China Normal University, Wuhan, China;National Language Resources Monitor Research Center for Network Media, Central China Normal University, Wuhan, China;School of Computer Science, Central China Normal University, Wuhan, China
摘要：
It is commonly known that food nutrition is closely related to human health. The complex interactions between food nutrients and diseases, influenced by gut microbial metabolism, present challenges in systematizing and practically applying knowledge. To address this, we propose a method for extracting triples from a vast amount of literature, which is used to construct a comprehensive knowledge graph on nutrition and human health. Concurrently, we develop a query-based question answering system over our knowledge graph, proficiently addressing three types of questions. The results show that our proposed model outperforms other state-of-art methods, achieving a precision of 0.92, a recall of 0.81, and an F1 score of 0.86in the nutrition and disease relation extraction task. Meanwhile, our question answering system achieves an accuracy of 0.68 and an F1 score of 0.61 on our benchmark dataset, showcasing competitiveness in practical scenarios. Furthermore, we design five independent experiments to assess the quality of the data structure in the knowledge graph, ensuring results characterized by high accuracy and interpretability. In conclusion, the construction of our knowledge graph shows significant promise in facilitating diet recommendations, enhancing patient care applications, and informing decision-making in clinical research.
期刊：
IEEE Transactions on Neural Networks and Learning Systems,2024年PP ISSN：2162-237X
通讯作者：
He, TT
作者机构：
[Fan, Rui] Cent China Normal Univ, Fac Artificial Intelligence Educ, Hubei Prov Key Lab Artificial Intelligence & Smart, Wuhan 430079, Peoples R China.;[Fan, Rui; Dong, Ming; Tu, Xinhui; Zhang, Mengyuan; He, Tingting; Chen, Menghan] Cent China Normal Univ, Natl Language Resources Monitor & Res Ctr Network, Wuhan 430079, Peoples R China.;[Dong, Ming; Tu, Xinhui; Zhang, Mengyuan; He, Tingting; Chen, Menghan] Cent China Normal Univ, Sch Comp, Hubei Prov Key Lab Artificial Intelligence & Smart, Wuhan 430079, Peoples R China.
通讯机构：
[He, TT ] C;Cent China Normal Univ, Natl Language Resources Monitor & Res Ctr Network, Wuhan 430079, Peoples R China.;Cent China Normal Univ, Sch Comp, Hubei Prov Key Lab Artificial Intelligence & Smart, Wuhan 430079, Peoples R China.
摘要：
Multimodal aspect-based sentiment classification (MABSC) aims to identify the sentiment polarity toward specific aspects in multimodal data. It has gained significant attention with the increasing use of social media platforms. Existing approaches primarily focus on analyzing the content of posts to predict sentiment. However, they often struggle with limited contextual information inherent in social media posts, hindering accurate sentiment detection. To overcome this issue, we propose a novel multimodal dual cause analysis (MDCA) method to track the underlying causes behind expressed sentiments. MDCA can provide additional reasoning cause (RC) and direct cause (DC) to explain why users express certain emotions, thus helping improve the accuracy of sentiment prediction. To develop a model with MDCA, we construct MABSC datasets with RC and DC by utilizing large language models (LLMs) and visual-language models. Subsequently, we devise a multitask learning framework that leverages the datasets with cause data to train a small generative model, which can generate RC and DC, and predict the sentiment assisted by these causes. Experimental results on MABSC benchmark datasets demonstrate that our MDCA model achieves the state-of-the-art performance, and the small fine-tuned model exhibits superior adaptability to MABSC compared to large models like ChatGPT and BLIP-2.
摘要：
Dialogue state tracking (DST) is a core component of task-oriented dialogue systems. Recent works focus mainly on end-to-end DST models that omit the spoken language understanding (SLU) module to directly obtain the dialogue state based on a user’s dialogue. However, the slot information detected by slot filling in SLU is closely tied to the slot–value pair that needs to be updated in DST. Efficient use of the key slot semantic knowledge obtained by slot filling contributes to improving the performance of DST. Based on this idea, we introduce slot filling as a subtask and build an end-to-end joint model to explicitly integrate the slot information detected by slot filling, which further guides DST. In this article, a novel stack-propagation framework with slot filling for multidomain DST is proposed. The stack-propagation framework is introduced to jointly model slot filling and DST. The framework directly feeds the key slot semantic knowledge detected by slot filling into the DST module. In addition, a slot-masked attention mechanism is designed to enable DST to focus on the key slot information obtained by slot filling. When the slot value is updated, a slot–value softcopy mechanism is designed to enhance the influence of the words marked by key slots. Experiments show that our approach outperforms previous methods and performs outstandingly on two benchmark datasets. IEEE
作者机构：
[Jing Hu; Po Hu] School of Computer Science, Central China Normal University, Wuhan, China;[Lingfei Wu] Pinterest, San Francisco, USA;[Yu Chen] Meta, Mountain View, USA;[Mohammed J.Zaki] Rensselaer Polytechnic Institute, Troy, USA
通讯机构：
[Lingfei Wu] P;Pinterest, San Francisco, USA
摘要：
WaveMLP has demonstrated remarkable performance in various vision tasks, such as dense feature detection and semantic segmentation. However, WaveMLP, as a local model, imposes limitations on fully connected layers by only allowing connections between tokens within the same local window. This constraint makes the model neglect the relationship among tokens in different windows, leading to a local token fusion and a degraded modeling performance. Specially, it poses challenges when dealing with hyperspectral image (HSI) classification tasks that require capturing long-range dependencies. To address this issue, this letter proposes a new position-aware WaveMLP, dubbed PA-WaveMLP, which incorporates a global polar positional encoding module (PPEM) into WaveMLP. PPEM is a lightweight method to encode the spatial relationship between land objects in distance and direction by using the radius and angle. By PPEM, the proposed PA-WaveMLP enables tokens to include their own spatial position information to the fusion process, allowing for the capture of long-range dependencies, while maintaining the excellent modeling capabilities of WaveMLP. The experimental results on three publicly available HSI datasets validate the effectiveness and generalizability of this newly proposed PA-WaveMLP. In particular, PA-WaveMLP model achieved an overall accuracy (OA) of 99.16%, 99.71%, and 99.47% on Indian Pines (IP), Pavia University (PU), and Salinas (SA), respectively.
摘要：
<jats:title>Abstract</jats:title>
<jats:p>With the development of information networks, the entities from different network domains interact with each other more and more frequently. Therefore, identity management and authentication are essential in cross-domain setting. The traditional Public Key Infrastructure (PKI) architecture has some problems, including single point of failure, inefficient certificate revocation status management and also lack of privacy protection, which cannot meet the demand of cross-domain identity authentication. Blockchain is suitable for multi-participant collaboration in multi-trust domain scenarios. In this paper, a cross-domain certificate management scheme CD-BCM based on the consortium blockchain is proposed. For the issue of Certificate Authority’s single point of failure, we design a multi-signature algorithm. In addition, we propose a unified structure for batch certificates verification and conversion, which improve the efficiency of erroneous certificate identification. Finally, by comparing with current related schemes, our scheme achieves good functionality and scalability in the scenario of cross-domain certificate management.</jats:p>
作者机构：
[Liu, Ming; E, Jinxuan; Liu, M; He, Chao] Cent China Normal Univ, Sch Comp Sci, Wuhan, Peoples R China.;[Liu, Rong] Cent China Normal Univ, Coll Phys Sci & Technol, Wuhan, Peoples R China.
会议名称：
5th International Conference on Computing, Networks and Internet of Things (CNIOT)
会议时间：
MAY 24-26, 2024
会议地点：
Tokyo, JAPAN
会议主办单位：
[He, Chao;E, Jinxuan;Liu, Ming] Cent China Normal Univ, Sch Comp Sci, Wuhan, Peoples R China.^[Liu, Rong] Cent China Normal Univ, Coll Phys Sci & Technol, Wuhan, Peoples R China.
摘要：
Virtual try-on is an image generation task for changing characters' clothes while preserving the characters' and the cloth's original attributes. Existing methods usually apply the traditional appearance flow method, which is susceptible to complex body postures or occlusions, leading to unclear texture of target clothing or distorted limbs of characters. We apply a StyleGAN-based(generative adversarial network) flow generator to estimate appearance flow, which provides more global information to overcome the issue. Additionally, more local information is captured to refine the appearance flow by adopting the channel attention mechanism. Qualitative and quantitative experiments demonstrate that our model can generate more realistic images.
