Forecasting 3D human motion is an important embodiment of fine-grained understanding and cognition of human behavior by artificial agents. Current approaches excessively rely on implicit network modeling of spatiotemporal relationships and motion characteristics, falling into the passive learning trap that results in redundant and monotonous 3D coordinate information acquisition while lacking actively guided explicit learning mechanisms. To overcome these issues, we propose an Active Perceptual Strategy (APS) for human motion prediction, leveraging quotient space representations to explicitly encode motion properties while introducing auxiliary learning objectives to strengthen spatio-temporal modeling. Specifically, we first design a data perception module that projects poses into the quotient space, decoupling motion geometry from coordinate redundancy. By jointly encoding tangent vectors and Grassmann projections, this module simultaneously achieves geometric dimension reduction, semantic decoupling, and dynamic constraint enforcement for effective motion pose characterization. Furthermore, we introduce a network perception module that actively learns spatio-temporal dependencies through restorative learning. This module deliberately masks specific joints or injects noise to construct auxiliary supervision signals. A dedicated auxiliary learning network is designed to actively adapt and learn from perturbed information. Notably, APS is model agnostic and can be integrated with different prediction models to enhance active perceptual.The experimental results demonstrate that our method achieves the new state-of-the-art, outperforming existing methods by large margins: 16.3% on H3.6M, 13.9% on CMU Mocap, and 10.1% on 3DPW.

Federated Unlearning (FU) has emerged as a promising paradigm for effectively removing the influence of specific data of clients from the global model in federated learning. It can further enhance personal data privacy for individual clients and eliminate the impact of malicious attacks like poisoning. Due to these benefits, many FU methods have been analyzed and proposed. Yet, they largely overlook external threats against FU systems, such as gradient inversion attacks that reconstruct client data from shared gradients, posing serious privacy risks to other participating clients. Motivated by this, we propose DIARY, a Differential prIvacy IntegrAted fedeRated recoverY framework to address these dual threats. DIARY presents a privacy budget allocation method, whose insight lies in adaptively assigning appropriate privacy budgets to various training and recovery phases to fully utilize the global privacy budget of each client, balancing the trade-off between privacy and utility. Furthermore, DIARY introduces a novel Federated noise-Immune aNomaly Detection (FIND) module. The deep integration of FIND with two-level selective storage and model rollback mechanisms contributes to model recovery, while significantly reducing the associated overhead. Finally, both rigorous theoretical analysis and extensive simulations compared with state-of-the-art methods are conducted to validate the effectiveness of DIARY.

With the rapid advancement of Internet of Things and artificial intelligence, device automation systems have become increasingly integrated with physical environments, introducing new security challenges for Trigger-Action Programming. An improper configuration of TAP rules may lead to severe cascading threats. However, existing methods typically rely on predefined safety properties and fail to capture the underlying semantic dependencies and interactions among rules. To address these limitations, we propose TAPGuard, a semantics-enhanced framework for TAP rule linkage modeling and cascading threat detection. Specifically, we identify two types of cascading threats: explicit threats, which arise from direct device interactions, and implicit threats, which are induced by shared environmental variables and may propagate across semantically related but structurally disconnected rules. TAPGuard leverages large language models to extract structured semantic elements from natural language rule descriptions and incorporates a semantic alignment module to assess the functional similarity between rules. Building on this, we propose a dual-relation context encoder incorporating node-level and semantic-level attention to model heterogeneous dependencies and enable multi-hop relational reasoning in the heterogeneous TAP rule graph. We evaluate TAPGuard on a real-world smart home dataset and demonstrate its effectiveness in detecting cascading threats. Experimental results show that TAPGuard significantly outperforms state-of-the-art graph-based baselines.

We propose NVLog, an NVM-based acceleration framework for disk file systems, designed to transparently harness the high performance of NVM within the legacy storage stack. NVLog provides on-demand byte-granularity sync absorption, reserving the fast DRAM path for asynchronous operations, meanwhile occupying NVM space only temporarily. To accomplish this, we designed a highly efficient log structure, developed mechanisms to address heterogeneous crash consistency, optimized for small writes, and implemented robust crash recovery and garbage collection methods. Compared to previous solutions, NVLog is lighter, more stable, and delivers higher performance, all while leveraging the mature kernel software stack and avoiding data migration overhead. Experimental results demonstrate that NVLog can accelerate disk file systems by up to 15.09x and outperform NOVA and SPFS in various scenarios by up to 3.72x and 324.11x, respectively.

Learning from label proportions (LLP), i.e., a challenging weakly-supervised learning task, aims to train a classifier by using bags of instances and the proportions of classes within bags, rather than annotated labels for each instance. Beyond the traditional bag-level loss, the mainstream methodology of LLP is to incorporate an auxiliary instance-level loss with pseudo-labels formed by predictions. Unfortunately, we empirically observed that the pseudo-labels are are often inaccurate due to over-smoothing, especially for the scenarios with large bag sizes, hurting the classifier induction. To alleviate this problem, we suggest a novel LLP method, namely Learning from Label Proportions with Auxiliary High-confident Instance-level Loss (L2P-AHIL). Specifically, we propose a dual entropy-based weight (DEW) method to adaptively measure the confidences of pseudo-labels. It simultaneously emphasizes accurate predictions at the bag level and avoids overly smoothed predictions. We then form high-confident instance-level loss with DEW, and jointly optimize it with the bag-level loss in a self-training manner. The experimental results on benchmark datasets show that L2P-AHIL can surpass the existing baseline methods, and the performance gain can be more significant as the bag size increases. The implementation of our method is available at https://github.com/TianhaoMa5/LLP-AHIL.

Quantum computers, leveraging superposition and entanglement, offer significant qubit efficiency for data processing compared to classical systems. However, encoding classical data into quantum states, given the current limitations of quantum hardware, often results in higher runtime complexity than classical methods, thus limiting the perceived quantum advantage. Previous quantum data compression methods, primarily based on Amplitude Encoding and mixed-state systems, result in lossy data recovery and necessitate extensive preprocessing. In this work, we propose Quantum Run-Length Encoding (QRLE), a novel lossless quantum data compression method that integrates Basic Encoding with Run-Length Encoding principles. By encoding repeated data sequences with their run lengths, QRLE achieves efficient and accurate data recovery on quantum computers, while exponentially reducing both qubit costs and runtime complexity compared to existing quantum data storage models. We further explore QRLE’s application in image processing, where it significantly optimizes quantum resource utilization over recent quantum image representation techniques. Experiments conducted on both quantum simulators and IBM’s superconducting quantum computer validate the efficiency of QRLE and confirm its compatibility with current quantum hardware.

Multi-label causal feature selection has garnered considerable attention for its ability to identify the most informative features while accounting for the causal dependencies between labels and features. However, previous work often overlooks the unique contributions of labels to the target variables in multi-label settings, focusing instead on prioritizing feature variables. Moreover, existing methods typically rely on traditional Markov Blanket (MB) discovery to construct an initial MB, which often fails to explore the most valuable form of spouse variables to feature selection in multi-label scenarios, leading to significant computational overhead due to redundant Conditional Independence (CI) tests required for spouse search. To address these challenges, we propose the Multi-label Causal Feature Selection Method with Optimal Spouses Discovery, MCF-Spouse, which leverages mutual information to quantify the contributions of both labels and features, ensuring the retention of the most informative variables in multi-label settings. Moreover, we systematically analyzes all potential forms of spouse variables to identify the optimal spouse case, significantly reducing the spouse search space and alleviating the time overhead associated with CI tests. Experiments conducted on diverse real-world datasets demonstrate that MCF-Spouse consistently outperforms state-of-the-art methods across multiple metrics, offering a scalable and interpretable solution for multi-label causal feature selection.

To address the limitations of existing clean-label backdoor attacks, particularly concerning feature space heterogeneity, we propose a novel feature-distance-guided clean-label backdoor attack method. Specifically, we first introduce a sample selection strategy based on feature discrepancy and recognizability constraints. This strategy involves calculating the mean feature vector for each class within the dataset and subsequently employing the Fréchet Inception Distance to quantify the deviation of individual sample feature vectors from their respective class mean, thereby identifying samples significantly diverging from the class distribution. Subsequently, we present an innovative noise generation technique termed Feature Displacement-Driven Noise Iteration. For each selected training sample, we iteratively adjust the intensity of the added noise to effectively amplify its feature distribution divergence while rigorously preserving the recognizability of the sample’s original label, ultimately yielding highly optimized adversarial noise. Finally, this iterated noise, along with a pre-defined trigger, is embedded into the chosen training samples to construct poisoned samples, which are then utilized for model training. Experimental results unequivocally demonstrate that, compared to existing techniques, our proposed method significantly enhances the attack success rate by 0.88% to 13.99%, while maintaining nearly identical classification accuracy on benign samples. This conclusively validates the superior performance of our approach.

Deep learning-based applications have great potential for providing intelligent and personalized services in the Internet of Things (IoT). However, the resource limitation in IoT devices may significantly hinder deep learning applications in IoTs, especially when infrastructures are absent for critical environments. Unmanned Aerial Vehicle (UAV) based split learning can alleviate this problem, by offloading the major part of the deep learning training tasks from IoT devices to the UAV. Whereas, current studies often overlook the latent privacy challenges caused by the UAV and extra data transmissions. To address this issue while ensuring efficient split learning, we propose a novel privacy-preserving split learning architecture. Based on this architecture, we present an improved pipeline scheme to synchronize the training and communicating period between the UAV and the IoT device. Then, in the context of privacy preservation, we formulate an optimization model to minimize the system energy consumption by jointly optimizing model split points, UAV service slot allocation, and flight trajectories. Based on Block Coordinate Descent (BCD) and Successive Convex Approximation (SCA), we put forward HOTSS algorithm to find the optimized solution of this model. Simulation results show the fluctuating characteristic of energy consumption changed with the increase of the privacy preservation requirement, and show our approach can reduce overall system energy consumption by an average of 6.7% compared to the benchmark scheme.

With the rise of blockchain technology, the trend toward decentralization has spread to the field of remote storage, leading to the emergence of decentralized storage as a promising model. This change is highlighted by its features of open and fair access, reduced dependence on intermediaries, and strong privacy protections. However, similar to centralized storage, the decentralization of data management presents challenges, including the separation of ownership and control, along with the need for integrity auditing on externally managed data. The current popular centralized auditing model for the mainstream cloud storage cannot be directly used for decentralized storage environments. Additionally, Homomorphic Verification Tag (HVT)-based auditing models encounter significant problems such as high computational costs and inefficient auditing processes. In response to these needs, we introduce a novel Byzantine-tolerant Lazy Auditing framework (BLA) to ensure data integrity in decentralized storage settings. A key innovation is the hierarchical architecture used: the upper level employs a simplified Practical Byzantine Fault Tolerance (PBFT) protocol to help nodes reach a consensus on data integrity audits. At the lower level, nodes are grouped into clusters based on criteria such as accessibility, organized using a block design strategy. This approach reduces unnecessary information exchange during the auditing process. It maximizes parallel processing and strengthens fault tolerance and system resilience. By distributing data, it also reduces the impact of node failures. Our theoretical analyses and empirical evaluations clearly show that BLA reduces communication complexity compared with conventional PBFT protocols. Additionally, when compared with traditional HVT-based schemes, BLA demonstrates better storage efficiency and improved computational performance, making it a viable and effective solution for data integrity auditing in decentralized storage systems.

Quantum computing offers parallel processing capabilities and resource-saving advantages, particularly useful for managing expansive datasets and complex image processing tasks. Grayscale images, being the simplest single-channel image mode, are frequently employed in artificial intelligence training. Before actual image applications, various image processing operations are typically required. However, the restoration of a grayscale image of dimensions 2n× 2nafter a series of linear transformations poses a challenge. Existing methods typically involve finding the inverse of the most recent linear transformation or re-encoding the image followed by repeated operations until the final transformation, resulting in excessive computational overhead and disconnection from subsequent quantum grayscale image applications. To address this issue, we propose a universal quantum linear restoration algorithm for grayscale image, denoted as QLR, which effectively bridges the stages of linear transformation and subsequent image applications. QLR reduces the time complexity from O(2n) to O(n) compared to classical counterpart. Building upon the QLR algorithm, we further propose two quantum resource-optimized compression methods for optional lossless image storage. Combining with other quantum algorithms and techniques, we design a framework (QGIP) aimed at bridging the processes of quantum grayscale image processing and applications. Experiments simulated on the IBM Quantum platform validate the correctness and efficiency of our proposal.

The increased utilization of Trigger-Action Programming (TAP) rules in smart homes has raised concerns regarding potential security threats in the interactions between smart digital devices/online services (DD/OS) and the physical environment. To ensure the secure use of intelligent and convenient infrastructure for users, we introduce an approach aimed at detecting potential security threats. In this paper, we propose IoT security threat models and categorize the threats into Risky DD/OS with Physical Security, Contradictory Operation of DD/OS and Environmental Impact Conflict. To effectively detect security threats, we construct an Internet of Things-Heterogeneous Information Networks (IoT-HIN) and enhance it with a knowledge base tool, transforming it into a knowledge-based IoT-HIN. We establish meta-paths to conduct analysis of events triggered by rules, and a threat detection algorithm is proposed to identify potential security threats and determine the rules leading to these threats. The proposed approach is validated using a real-world dataset, and the experimental results demonstrate its efficiency and practicality. Furthermore, a comparative analysis with similar works is conducted to highlight the superiority of our proposed approach.

The rapid expansion of the Internet of Things (IoT) has led to the development of smart homes and automation systems. Trigger-action programming (TAP) has emerged as a prevalent paradigm used in IoT, facilitating the creation of automation rules. However, with the proliferation of TAP rules, the potential for conflicts between them grows significantly, which results in undesired outcomes or even safety risks. In this article, we propose a context-aware conflict detection framework for TAP rules, called CCDF-TAP, to identify the potential rule conflicts. Specifically, we incorporate external knowledge and context information during the TAP data preprocessing stage, which is conducive to accurately defining the rule conflicts. Then, based on the above information, the conflict types are defined and a conflict graph is constructed, which establishes a unified format for the rule conflict detection task. Finally, we propose a novel algorithm called dual-channel graph attention auto-encoders (DualGAAs) for efficient conflict detection, which takes the conflict graph as the input and excels in accurately identifying conflicts. Extensive experiments conducted on a comprehensive IFTTT data set demonstrate the superiority of DualGAA in detecting conflicts, achieving an exceptional accuracy of 98.85% and an F1 score of 98.91%. The contributions of our study offer a comprehensive end-to-end solution for context-aware conflict detection in TAP rules, thereby significantly enhancing the security and dependability of IoT smart home systems.

Continuous and remote health monitoring medical applications with heterogeneous requirements can be realized through wireless body area networks (WBANs). Energy harvesting is adopted to enable low-power health applications and long-term monitoring without battery replacement, which have drawn significant interest recently. Because energy harvesting WBANs are obviously different from battery-powered ones, network protocols should be designed accordingly to improve network performance. In this article, an efficient cross-layer media access control protocol is proposed for radio frequency powered energy harvesting WBANs. We redesigned the superframe structure, which can be rescheduled by the coordinator dynamically. A time switching (TS) strategy is used when sensors harvest energy from radio frequency signals broadcast by the coordinator, and a transmission power adjustment scheme is proposed for sensors based on the energy harvesting efficiency and the network environment. Energy efficiency can be effectively improved that more packets can be uploaded using limited energy. The length of the energy harvesting period is determined by the coordinator to balance the channel resources and energy requirements of sensors and further improve the network performance. Numerical simulation results show that our protocol can provide superior system performance for long-term periodic health monitoring applications.