Neurons exert a pivotal role in the preservation of cardiac physiological function. However, there is a lack of explanation about the mechanism of cardiac neurons in the pathogenesis of cardiac dysfunction. Here, we generated a cardiac neuron landscape including 11,026 neuronal cells based on the integration of published single-nucleus RNA sequencing data from 75 patients with heart failure and 45 healthy donors. We determined ten distinct neuronal cell subsets differing in abundances, compositions, and biological functions in the heart. In particular, N4-ALK neurons were significantly enriched in failing hearts relative to healthy controls, and their abundance was associated with the response to left ventricular assist device implantation. RXRG, a transcription factor highly expressed in neuronal cells, participated in the transcriptional regulatory network of N4-ALK neurons and showed a positive correlation with the expression of their marker genes. Notably, in heart failure, the PTN-PTPRZ1 axis mediated specific crosstalk between cardiac fibroblasts and N4-ALK neurons. Finally, we used N4-ALK-related features to develop an optimized prediction model for identifying individuals with heart failure. Overall, our integrative cardiac neuron atlas comprehensively characterizes the molecular and functional diversity of neuronal cells, providing a new perspective for further exploration of the regulatory function of neurons in heart failure.

Recent years have witnessed the notable resurgence of graph neural networks (GNNs) in service recommender systems. GNN-based service recommender systems heavily rely on the recursive message propagation mechanism between layers. However, they encounter challenges due to data sparsity and susceptibility to noisy interactions, leading to unstable and suboptimal performance. Several studies have attempted to mitigate these issues by integrating GNNs with contrastive learning techniques. Despite their success, most existing approaches utilize either stochastic or heuristic data augmentation methods. We argue that these methods suffer from inherent limitations, including disruption of the original structure of the user-service interaction graph and the need for laborious and iterative experimentation. In this paper, we propose a novel Diffusion model-based Graph Contrastive Learning framework, named DiffGCL, for service recommendation. Specifically, we incorporate Diffusion models into service recommender systems as generators for contrastive learning views. We also introduce a self-adaptive data augmentation strategy that dynamically constructs new dual views each epoch while preserving the intrinsic semantic structure of the generated graph. Extensive experiments conducted on multiple datasets demonstrate the effectiveness of DiffGCL, outperforming various baseline models on service recommendation tasks.

Cloud platforms provide extendable and flexible orchestration capabilities for microservices by consolidating multiple computing resources, e.g., multiple service instances can be deployed redundantly to form server clusters, so that if a cloud node or service instance fails unexpectedly, a copy of the service on another node can take over the failed instance to ensure persistent reliability of the system. However, such redundant deployments in cloud may result in unbalanced utilisation of resources, which complicates software quality assessment on cloud nodes. Moreover, the maximisation of the quality of service (QoS) is recognised as a NP problem, making cloud services difficult to be optimised in real-time. To solve these problems, this paper proposes a scheduling combination method for microservices on cloud platforms based on a dual-layer QoS evaluation model, which renders the dual superposition effect of cloud nodes and service instances on system quality. The advantage of the model is that it considers the coupling relationship between cloud hardware or software quality. Further, to optimise QoS in cloud environments based on the model, a hybrid Vision-improved Ant Colony-Genetic algorithm called VACG is proposed to solve the combination optimisation problem and find optimum composition solutions. Our experimental results demonstrate that the scheduling policy based on the dual-layer QoS model surpasses another non-scheduling policy on the metrics of system QoS index and service level agreement conflicts. Additionally, it is found that VACG has a 17.13% and 27.03% of improvement on the optimisation accuracy over a genetic algorithm and ant colony optimisation respectively, as well as higher computational efficiency and stability.

Microservices architecture is a new paradigm for application development.The problem of optimising the performance of microservice architectures from a non-functional perspective is a typical Nondeterministic Polynomial (NP) problem.Therefore, aiming to quantify the non-functional requirements of computing microservice systems, while solving the problem of latency in computing the best combination of services with the maximum QoS objective function value, this paper proposes a microservice combination approach based on the QoS model and a CGWO algorithm for optimisation computation for this model.The experimental results verify that the error rate of the method is only 0.528% on the non-functional combination optimisation problem, and the computational efficiency of the algorithm increases by 97.29% when the complexity of the problem search space increases, while CGWO improves 65.97% and 81.25% respectively in the accuracy of optimisation compared to the prototype of the algorithm (GWO), and has a stable optimisation performance, aspect.It proves that the research in this paper has a high advantage in automatically searching for the best QoS for the microservice combination problem.

Recently low-bit (e.g., 8-bit) network quantization has been extensively studied to accelerate the inference. Besides inference, low-bit training with quantized gradients can further bring more considerable acceleration, since the backward process is often computation-intensive. Unfortunately, the inappropriate quantization of backward propagation usually makes the training unstable and even crash. There lacks a successful unified low-bit training framework that can support diverse networks on various tasks. In this paper, we give an attempt to build a unified 8-bit (INT8) training framework for common convolutional neural networks from the aspects of both accuracy and speed. First, we empirically find the four distinctive characteristics of gradients, which provide us insightful clues for gradient quantization. Then, we theoretically give an in-depth analysis of the convergence bound and derive two principles for stable INT8 training. Finally, we propose two universal techniques, including Direction Sensitive Gradient Clipping that reduces the direction deviation of gradients and Deviation Counteractive Learning Rate Scaling that avoids illegal gradient update along the wrong direction. The experiments show that our unified solution promises accurate and efficient INT8 training for a variety of networks and tasks, including MobileNetV2, InceptionV3 and object detection that prior studies have never succeeded. Moreover, it enjoys a strong flexibility to run on off-the-shelf hardware, and reduces the training time by 22% on Pascal GPU without too much optimization effort. We believe that this pioneering study will help lead the community towards a fully unified INT8 training for convolutional neural networks.