Huahu Xu

With the large-scale deployment of industrial cyber-physical system (ICPS) terminals such as drones and robots, edge computing often falls short in multidevice collaboration scenarios. Fog computing overcomes this issue by integrating different resources across a distributed node network, thereby enabling more efficient system-wide coordination. However, the limited resources of fog nodes make independently addressing the efficient scheduling demands of massive devices and complex tasks difficult. To address this problem, an attention-enhanced proximal policy optimization (APPO) algorithm is proposed in this paper. First, a dual-branch attention-based architecture is designed to model the complex dependencies between task requirements and multilayer resource states, thereby effectively capturing the coupling relationships in a hierarchical environment. Next, the PPO framework is leveraged with a centralized critic that integrates global system views, which enables coordinated scheduling decisions that are optimized across a latency-sensitive fog layer and a resource-rich cloud layer. Finally, a fault tolerance strategy is constructed to address the higher failure probabilities of distributed fog nodes through real-time monitoring and dynamic rescheduling, which significantly increases the robustness of the system. The experimental results demonstrate that the proposed APPO algorithm outperforms the existing baseline approaches and achieves a 28.15% higher task success rate while reducing virtual machine (VM) costs, energy consumption, and average delays by 20.94%, 21.33%, and 56.12%, respectively.

As sensor-cloud networks scale to support the growing demands of industrial Internet of Things (IIoT) applications, efficient virtual machine (VM) scheduling becomes essential for optimizing both energy efficiency and quality of service (QoS). In these environments, the collaboration between large-scale and lightweight models deployed across cloud resources introduces significant challenges in dynamic resource allocation, task scheduling, and energy-aware computing scenarios. A novel reinforcement learning-guided Monte Carlo VM scheduling (RMVS) method, which integrates Monte Carlo tree search (MCTS) and deep reinforcement learning (DRL) to enable intelligent and adaptive resource management, is proposed in this paper. The DRL model is used to provide prior knowledge, guiding the MCTS to efficiently explore candidate physical machines (PMs) and make better scheduling decisions. Notably, the integration of VM migration allows the system to dynamically adjust workloads, real-locate resources, and shut down idle servers, thus improving both the energy efficiency and CPU utilization. The proposed RMVS framework demonstrates significant improvements, reducing the makespan, energy consumption, and Service Level Agreement (SLA) violation rate by approximately 4.74%, 4.34%, and 11.67%, respectively, while improving CPU utilization by about 6.23%, outperforming state-of-the-art methods.

Cranial imaging diagnosis analyzes lesions in cranial CT images, automatically determining information such as lesion type, providing physicians with diagnostic recommendations, and saving valuable time during urgent medical events like stroke. Most current mainstream diagnostic generation methods rely on single-modal data for analysis, such as imaging reports or a single CT slice. However, relying on a single CT slice leads to loss of lesion information which across different CT slices. In addition, imaging reports often contain only coarse descriptions of abnormal regions without detailed pixel-level features, resulting in insufficient lesion characterization. Moreover, some works directly use 3D CNN to process entire 3D CT scans, but using only 3D CNN makes it challenging to accurately identify very small lesions. Therefore, this paper proposes a multimodal diagnostic model called MM-CD (MultiModal Craniocerebral Diagnose), which integrates imaging reports and cranial 3D CT findings for joint diagnosis. Specifically, this study first uses a 2D image pretrained model in combination with a vertical-dimension weight generation module for the cranial 3D CT images, enabling the model to focus on abnormal CT slices. Next, a multi-scale image fusion module is designed to effectively consolidates lesion descriptions from multiple CT slices into a single slice. Followed by a self-attention mechanism, this paper integrates CT information with the imaging report to construct a more comprehensive diagnostic reference. This clinically oriented design aims to lower missed-diagnosis rates for small, spatially sparse lesions and to shorten door-to-treatment intervals in acute care. Experimental results on a real clinical dataset show that the method improves overall accuracy by 1.65% compared to existing state-of-the-art medical multimodal models.

Video Moment Retrieval (VMR) is a key cross-modal task with broad theoretical and practical applications. While fully supervised methods deliver strong performance, they are constrained by the high cost of temporal boundary annotations. Weakly supervised methods mitigate this issue but suffer from limited accuracy due to coarse supervision. Recently, point-supervised approaches that leverage single-frame annotations as a cost-effective alternative have emerged as a promising paradigm. However, these methods often fail to leverage annotated frames for cross-modal semantic alignment. Additionally, they overlook global video structures and hierarchical segment relationships, leading to suboptimal retrieval accuracy under sparse supervision. To address these challenges, we propose the Adaptive Dual-Stage Tree Construction (ADTC) model, a novel framework designed specifically for point-supervised VMR. First, the model introduces a dual-stage hypothesis tree architecture that seamlessly integrates local and global trees, enabling the effective modeling of semantic relationships across multiple temporal scales. Second, it incorporates frame clustering and scene segmentation to extract the structural characteristics of video content, providing a foundation for comprehensive node relevance evaluation and an adaptive merging control strategy to optimize tree construction. Third, a hierarchical adaptive tree pruning strategy is implemented, combined with a novel proposal selection mechanism for distinguishing between positive and negative samples. These components are jointly optimized through a multilevel loss function, enabling enhanced semantic alignment and retrieval performance. The experimental results demonstrate that ADTC achieves state-of-the-art performance on the Charades-STA and ActivityNet Captions datasets under the point-supervised setting. On Charades-STA, it attains R@1 scores of 50.28% at IoU = 0.5 and 34.79% at IoU = 0.7, outperforming existing point-supervised methods. On ActivityNet Captions, it achieves R@1 scores of 65.02% at IoU = 0.3 and 46.13% at IoU = 0.5, establishing new benchmarks. Notably, it approaches fully supervised performance while significantly reducing annotation costs. Ablation studies further confirm the effectiveness of each model component.