Knowledge graph reasoning (KGR) is a key technology that infers missing facts in knowledge graphs (KGs). Given that real-world scenarios typically encounter unseen KGs with new entities and new relations, researchers have begun to explore fully inductive KGR methods. This setting presents greater challenges and has not been fully explored. Current methods primarily construct relation graphs based on the original KG to facilitate message passing between relations. These models have made significant progress in achieving fully inductive reasoning. However, as relation graphs focus solely on the co-occurrence patterns between relations, they often fail to capture reasoning patterns in KGs, which causes the model to struggle in effectively distinguishing between different relations and entities. This limitation severely restrict the reasoning capabilities of existing methods. In light of this, we propose the Adaptive Rule Pattern Transfer model (AdaRPT) for KGR. It aims to leverage logical rules for each relation in the KG to learn more comprehensive and transferable knowledge representations for entities and relations. For entities, we design a non-parameter message passing model that aggregates path information from the query entity to other entities. The path information for each entity is then matched with rules to obtain the transferable feature of each entity. And for relations, we extract both reasoning and co-occurrence patterns from KGs as transferable relation features. Finally, a path-based graph neural network (GNN) is employed on the transferable features of entities and relations to perform reasoning on KGs. Extensive experimental evaluations on 43 datasets for both inductive and transductive reasoning demonstrate the effectiveness and generalization capability of AdaRPT.

ABSTRACT Imbalanced data classification is one of the challenges in the field of data mining and machine learning. At present, the main method to solve imbalanced data classification issues from the data level is resampling. Hybrid sampling is widely used because it can avoid the problem of overfitting or mistakenly deleting the most useful samples when using oversampling or undersampling alone. However, the current hybrid sampling methods are mostly implemented in serial, which has the problems of excessive time cost and mutual influence. Meanwhile, few studies consider the automatic determination of the oversampling rate and undersampling rate in hybrid sampling methods but use the default sampling rate. Therefore, this study proposes a novel optimal sampling rate selection for a parallel hybrid sampling framework of imbalanced data classification. At the same time, we improved the differential evolution algorithm to optimize the sampling rate for the parallel hybrid sampling framework to obtain more stable classification performance. Experiments show that the SRDE specifically designed for the parallel hybrid sampling framework is superior to other heuristic algorithms on imbalanced data classification, and the results of statistical test experiments also verified this result.

Abstract Electronic health information exchange (HIE) allows doctors, nurses, pharmacists, other health care providers and patients to appropriately access and securely share a patient’s vital medical information electronically—improving the speed, quality, safety and cost of patient care. At present, public key cryptography (PKC) is the most secure and practical cryptographic techniques to achieve this function in healthcare information exchange for medical IoT systems. ElGamal cryptosystem is one of the most well-known public key cryptosystems (they are also called asymmetric key cryptosystems), which is based on the discrete logarithm problem. At present, with the development of quantum computer technology, the ElGamal cryptosystem may be attacked by quantum algorithms. At the same time, since ElGamal requires several secure random integers to resist cryptographic analysis and ensure communication security and securing data sharing. Especially, when the encrypted information is large, multiple random numbers need to be used for grouping encryption, which makes the efficiency of ElGamal need to be improved. It is necessary to construct an alternative cryptographic algorithm that is more secure and efficient than ElGamal. In this paper, we construct a new public key cryptosystem (PKC) based on the discrete logarithm problem in $$GL\left(n,p\right)$$ G L n , p , which is constructed by the invertible matrix multiplication and can become an alternative version of the ElGamal public key cryptosystem. We call it Matrix ElGamal cryptosystem (M-EPKC). It is proved that the proposed PKC is computationally secure, which can provide the same security as the ElGamal cryptosystem in a much smaller finite field and use fewer random integers when encrypting large amounts of messages. For matrices of size $$n$$ n , ElGamal PKC requires $$n$$ n times more random numbers to encrypt plaintext with the same amount of data. The proposed M-EPKC is not only proved to be resistant to Shor’s algorithm attack (Shor in SIAM Rev 41: 303–332, 1999) on the integer field, but it can also improve its own computational efficiency by accelerating the decryption algorithm. Therefore, compare with the ElGamal cryptosystem, our proposed M-EPKC can provide a more secure and efficient method in healthcare information exchange for medical IoT systems.

As the field of machine reading comprehension (MRC) continues to evolve, it is unlocking enormous potential for its practical application. However, the currently well-performing models predominantly rely on massive pre-trained language models with at least several hundred million or even over one hundred billion parameters. These complex models not only require immense computational power but also extensive storage, presenting challenges for resource-limited environments such as online education.Current research indicates that specific capabilities of larger models can be transferred to smaller models through knowledge distillation. However, prior to our work, there were no small models specifically designed for MRC task with complex reasoning abilities. In light of this, we present a novel multi-teacher multi-stage distillation approach, MTMS. It facilitates the easier deployment of reasoning-based MRC task on resource-constrained devices, thereby enabling effective applications. In this method, we design a multi-teacher distillation framework that includes both a logical teacher and a semantic teacher. This framework allows MTMS to simultaneously extract features from different perspectives of the text, mitigating the limitations inherent in single-teacher information representations. Furthermore, we introduce a multi-stage contrastive learning strategy. Through this strategy, the student model can progressively align with the teacher models, effectively bridging the gap between them. Extensive experimental outcomes on two inference-based datasets from real-world scenarios demonstrate that MTMS requires nearly 10 times fewer parameters compared with the teacher model size while achieving the competitive performance.

Abstract As mobile internet and Internet of Things technologies continue to advance, the application scenarios of peer-to-peer Internet of Drones (IoD) are becoming increasingly diverse. However, the development of IoD also faces significant challenges, such as security, privacy protection, and limited computing power, which require technological innovation to overcome. For group secure communication, it is necessary to provide two basic services, user authentication and group key agreement. Due to the limited storage of IoD devices, group key negotiation requires lightweight calculations, and conventional schemes cannot satisfy the requirements of group communication in the IoD. To this end, a new lightweight communication scheme based on ring neighbors is presented in this paper for IoD, which not only realizes the identity verification of user and group key negotiation, but also improves computational efficiency on each group member side. A detailed security analysis substantiates that the designed scheme is capable of withstanding attacks from both internal and external adversaries while satisfying all defined security requirements. More importantly, in our proposal, the computational cost on the user side remains unaffected by the variability of the number of members participating in group communication, as members communicate in a non-interactive manner through broadcasting. As a result, the protocol proposed in this article demonstrates lower computational and communication costs in comparison to other cryptographic schemes. Hence, this proposal presents a more appealing approach to lightweight group key agreement protocol with user authentication for application in the IoD.

The task of machine reading comprehension (MRC) is to enable machine to read and understand a piece of text and then answer the corresponding question correctly. This task requires machine to not only be able to perform semantic understanding but also possess logical reasoning capabilities. Just like human reading, it involves thinking about the text from two interacting perspectives of semantics and logic. However, previous methods based on reading comprehension either consider only the logical structure of the text or only the semantic structure of the text and cannot simultaneously balance semantic understanding and logical reasoning. This single form of reasoning cannot make the machine fully understand the meaning of the text. Additionally, the issue of sparsity in composition presents a significant challenge for models that rely on graph-based reasoning. To this end, a cross-graph knowledge propagation network (CGKPN) with adaptive connection is presented to address the above issues. The model first performs self-view node embedding on the constructed logical graph and semantic graph to update the representations of the graphs. Specifically, a relevance matrix between nodes is introduced to adaptively adjust node connections in response to the challenge posed by sparse graph. Subsequently, CGKPN conducts cross-graph knowledge propagation on nodes that are identical in both graphs, effectively resolving conflicts arising from identical nodes in different views, and enabling the model to better integrate the logical and semantic relationships of the text through efficient interaction. Experiments on the two MRC datasets ReClor and LogiQA indicate the superior performance of our proposed model CGKPN compared to other existing baselines.

With the rapid developments of Internet of Things (IoT) technologies, the security of sensitive data has attracted more and more attention for many resource-asymmetric smart environments, such as smart home, smart agriculture and so on. The resource-asymmetry environment refers to the uneven distribution of resources on different devices side, which is specifically manifested as gateway side is resource-rich, user side and device side are resource-restricted. Hence, a secure and practical authentication key establishment scheme for such smart environments is urgently needed. Recently many researchers have designed authentication and key establishment schemes for security purpose, however most of them cannot consider the excess of gateway resources and guarantee the anonymity of user, and further, they are not suitable for resource-asymmetric smart environments because they are not lightweight enough in user side and smart device side. Due to the fact that Rabin cryptosystem has the large difference in time-consuming between encryption and decryption, it is extremely suitable for constructing authentication and key establishment scheme for resource-asymmetric smart environments. So, a new practical authentication and key establishment scheme based on the Rabin cryptosystem for resource-asymmetric smart environments is proposed, which can make better use of the advantages of abundant gateway resources and realize the lightweight operations on device side and user side, and at the same time can provide user anonymity. With Proverif and BAN logic, we can prove that our solution not only provides anonymity, but also satisfies all defined security features. Simultaneously, compared with latest similar protocols in computation cost and communication overhead, the results show that our scheme is more effective. Hence, our design has more attraction for authentication and key establishment scheme in resource-asymmetric smart environments.

Pathology images contain rich information of cell appearance, microenvironment, and topology features for cancer analysis and diagnosis. Among such features, topology becomes increasingly important in analysis for cancer immunotherapy. By analyzing geometric and hierarchically structured cell distribution topology, oncologists can identify densely-packed and cancer-relevant cell communities (CCs) for making decisions. Compared to commonly-used pixel-level Convolution Neural Network (CNN) features and cell-instance-level Graph Neural Network (GNN) features, CC topology features are at a higher level of granularity and geometry. However, topological features have not been well exploited by recent deep learning (DL) methods for pathology image classification due to lack of effective topological descriptors for cell distribution and gathering patterns. In this paper, inspired by clinical practice, we analyze and classify pathology images by comprehensively learning cell appearance, microenvironment, and topology in a fine-to-coarse manner. To describe and exploit topology, we design Cell Community Forest (CCF), a novel graph that represents the hierarchical formulation process of big-sparse CCs from small-dense CCs. Using CCF as a new geometric topological descriptor of tumor cells in pathology images, we propose CCF-GNN, a GNN model that successively aggregates heterogeneous features (e.g., appearance, microenvironment) from cell-instance-level, cell-community-level, into image-level for pathology image classification. Extensive cross-validation experiments show that our method significantly outperforms alternative methods on H&E-stained and immunofluorescence images for disease grading tasks with multiple cancer types. Our proposed CCF-GNN establishes a new topological data analysis (TDA) based method, which facilitates integrating multi-level heterogeneous features of point clouds (e.g., for cells) into a unified DL framework.

With the increasing popularity and wide application of the Internet, the users (such as managers and data consumers) in the Industrial Internet of Things (IIoT) can remotely analyze and control real-time data collected by various smart sensor devices. However, there are many security and privacy issues in the process of transmitting collected data through public channels in IIoT environment. In order to against the illegal access by opponents, a novel anonymous user authentication and key agreement scheme based on hash and elliptic curve encryption is proposed in this article, which not only uses a pseudonym tuple database in control nodes to realize the functions of user dynamic joining and anonymity protection, but also resists key loss and device capture attacks through fuzzy biometric extraction technology. In addition, the formal secure analysis of the proposed scheme is carried out using the BAN logic model and ROR model, which proves the security of the proposed scheme. Meanwhile, we also prove the scheme can against the described existing attacks and meet the design goals by a detailed informal security discussion. Compared with the latest similar IIoT authentication proposals, our solution has a very obvious advantage in communication efficiency and realizes more functions. Hence, our scheme is more suitable for the IIoT environment, and can also generate greater benefits.

Abstract Group‐oriented applications show its potential ability in the next generation of wireless sensor networks (5G WSNs), which have the particularity of being heterogeneous and so have different capabilities in terms of storage, computing, communicating and energy. One of the main challenges for secure group‐oriented applications (SGA) in 5G WSNs is how to secure communication between these heterogeneous devices. Conventional protocols are not suitable for SGA in 5G sensor networks since multiparty output establishment in this environment requires lightweight communication and computation overhead, further the primary task of SGA in 5G WSNs is to securely transmit various types of jointly computing data. Hence, membership authentication and multiparty output for arithmetic computations become two fundamental and necessary security services in SGA for 5G WSNs. In this paper we propose a novel design of non‐interactive integrated membership authenticated multiparty output for arithmetic computations in 5G sensor networks, which embeds the function of membership authentication and multiparty output for arithmetic computations. Since any arithmetic computation function is composed of multiple additions and multiplications, our result serves as a general method for multiparty computation output in SGA. This design is more suitable for lightweight membership authenticated multiparty arithmetic computations output in 5G sensor networks.

Internet of Medical Things (IoMT) is a kind of Internet of Things (IoT) that includes patients and medical sensors. Patients can share real‐time medical data collected in IoMT with medical professionals. This enables medical professionals to provide patients with efficient medical services. Due to the high efficiency of cloud computing, patients prefer to share gathering medical information using cloud servers. However, sharing medical data on the cloud server will cause security issues, because these data involve the privacy of patients. Although recently many researchers have designed data sharing schemes in medical domain for security purpose, most of them cannot guarantee the anonymity of patients and provide access control for shared health data, and further, they are not lightweight enough for IoMT. Due to these security and efficiency issues, a novel lightweight privacy‐preserving data sharing scheme is constructed in this paper for IoMT. This scheme can achieve the anonymity of patients and access control of shared medical data. At the same time, it satisfies all described security features. In addition, this scheme can achieve lightweight computations by using elliptic curve cryptography (ECC), XOR operations, and hash function. Furthermore, performance evaluation demonstrates that the proposed scheme takes less computation cost through comparison with similar solutions. Therefore, it is fairly an attractive solution for efficient and secure data sharing in IoMT.

Cell nucleus segmentation is a fundamental task in biomedical image analysis. Generating realistic cell nucleus data with ground truth masks can help tackle difficulties such as insufficient training data for deep learning models and the need to deal with "hard" cases (e.g., tightly clumped nuclei). Known nucleus generation methods generated individual nucleus masks from parametric models or based on direct transformations of real masks. It is difficult for these methods to capture and simulate the distributions of real nuclei and interactions among hard nuclei. In this paper, we propose a new three-stage coarse-to-fine nucleus generation method for 2D and 3D nucleus segmentation. The first stage simulates the positions and sizes of nuclei; the second stage simulates the shapes of nuclei and interactions among clumped nuclei; the third stage simulates the textures of nuclei. We evaluate our method on 2D and 3D cell nucleus image datasets. Experimental results show that our new nucleus generation method considerably helps improve cell nucleus segmentation performance and outperforms known nucleus generation methods for nucleus segmentation with a small amount of training data.

Deep learning has been applied successfully to many biomedical image segmentation tasks. However, due to the diversity and complexity of biomedical image data, manual annotation for training common deep learning models is very timeconsuming and labor-intensive, especially because normally only biomedical experts can annotate image data well. Human experts are often involved in a long and iterative process of annotation, as in active learning type annotation schemes. In this paper, we propose representative annotation (RA), a new deep learning framework for reducing annotation effort in biomedical image segmentation. RA uses unsupervised networks for feature extraction and selects representative image patches for annotation in the latent space of learned feature descriptors, which implicitly characterizes the underlying data while minimizing redundancy. A fully convolutional network (FCN) is then trained using the annotated selected image patches for image segmentation. Our RA scheme offers three compelling advantages: (1) It leverages the ability of deep neural networks to learn better representations of image data; (2) it performs one-shot selection for manual annotation and frees annotators from the iterative process of common active learning based annotation schemes; (3) it can be deployed to 3D images with simple extensions. We evaluate our RA approach using three datasets (two 2D and one 3D) and show our framework yields competitive segmentation results comparing with state-of-the-art methods.

3D image segmentation plays an important role in biomedical image analysis. Many 2D and 3D deep learning models have achieved state-of-the-art segmentation performance on 3D biomedical image datasets. Yet, 2D and 3D models have their own strengths and weaknesses, and by unifying them together, one may be able to achieve more accurate results. In this paper, we propose a new ensemble learning framework for 3D biomedical image segmentation that combines the merits of 2D and 3D models. First, we develop a fully convolutional network based meta-learner to learn how to improve the results from 2D and 3D models (base-learners). Then, to minimize over-fitting for our sophisticated meta-learner, we devise a new training method that uses the results of the baselearners as multiple versions of “ground truths”. Furthermore, since our new meta-learner training scheme does not depend on manual annotation, it can utilize abundant unlabeled 3D image data to further improve the model. Extensive experiments on two public datasets (the HVSMR 2016 Challenge dataset and the mouse piriform cortex dataset) show that our approach is effective under fully-supervised, semisupervised, and transductive settings, and attains superior performance over state-of-the-art image segmentation methods.

In this study, a novel interest point detection algorithm which combines image intensity variation and edge contour information is proposed. Firstly, the Canny edge contour detector is used to extract the edge map. Secondly, the imaginary parts of multi‐scale Gabor filters are applied to smooth input image and then the normalised information entropies at various scales can be acquired. Finally, multiplication of different normalised information entropies will be served as a new measure for interest point that can also be used for interest point detection. This method has two advantages: on the one hand, detection accuracy is greatly improved because combination information is adopted to extract interest points including contour shape information, grey variation of edge pixels and their neighbours; on the other hand, non‐interest points can be well inhabited due to multi‐scale product in detector is served as an interest point measure. Also, desirable noise robustness and time efficiency are validated through experiments. Compared with four other state‐of‐art methods, the proposed method shows excellent performance in terms of geometric transformations and localisation accuracy of repeated interest points.

This paper presents a statistical analysis of several key factors in H.264/AVC motion estimation, including the relationship between sum of absolute difference and sum of absolute transformed difference (SATD) in macroblock (MB) level, the changes of SATD from full-pixel precision to quarter-pixel precision, and so on. Based on these statistical characteristics, we proved the validity of a fractional motion estimation (FME) mode reduction criterion. During motion estimation, any MBs satisfying this criterion can keep their integer motion estimation best mode until after FME, thus avoiding many costly less probable prediction modes in the FME stage. Simulation results confirm that prediction schemes using this criterion have a very low miss rate in mode determination, FME speed-up is achieved and the peak signal-to-noise ratio drop is almost negligible.

In this paper, a highly efficient parallel algorithm of H.264 encoder is proposed. This algorithm fully exploits all data dependencies for maximal compression and the best reconstructed quality. Furthermore, this new method achieves the optimal compression at a frame rate that increases approximately linearly as the number of parallel processing elements (Z. Zhao and P. Liang). Both the encoding speed and compression efficiency are improved compared to previous approaches. This paper also gives the relation between the number of parallel processing elements and the theoretical encoding time and the relation between the number of processors and the number of concurrently processed frames. Our simulation results show that this algorithm can also achieve the optimal encoding quality as a sequential processing encoder provided by Joint Video Team (JVT)

This paper presents a new method for parallel processing of H.264 video encoder using data partition and task scheduling that fully exploits all data dependencies for maximal compression (Zhao and Liang). The new method achieves the optimal compression at a frame rate that increases approximately linearly as the number of parallel processing elements. This is a significant improvement over prior art parallel encoders for H.264 which invariably sacrifice data dependency and/or optimal coding mode. It is shown that the new method outperforms prior approaches in both encoding speed and compression efficiency. This paper also gives the relation between the number of parallel processing elements and the theoretical encoding time and the relation between the number of processors and the number of concurrently processed frames. Software simulation shows that this parallel processing method achieves the same compression quality as a sequential processing encoder, e.g., the JM series

The H.264/AVC video compression standard exhibits excellent compression ratio due to many new features compared with its previous counterparts. Among these features, multiple reference inter-prediction and variable block-size play very important roles. However, their implementations require high memory access bandwidth. For real-time implementation, it is critical to reduce the memory access bandwidth. For this purpose, this paper presents a Partial Forward Processing Algorithm (PFPA) for frame-level data re-use and a mode decision strategy. Simulation results show that up to 2/3 of the original memory access bandwidth can be saved when using five reference frames with a search range of [-64,64]. The new mode decision method avoids sharply increased memory size used to store intermediate processing results and at the same time, the PSNR for the decoded sequence is very close to the optimal result produced by the reference software JM9.0 provided by the Joint Video Team (JVT) [1].