Search

The primary challenge in Federated Learning (FL) is to model non-IID distributions across clients, whose fine-grained structure is important to improve knowledge sharing. For example, some knowledge is globally shared across all clients, some is only transferable within a subgroup of clients, and some are client-specific. To capture and exploit this structure, we train models organized in a multi-level structure, called ``Multi-level Additive Models (MAM)'', for better knowledge-sharing across heterogeneous clients and their personalization. In federated MAM (FeMAM), each client is assigned to at most one model per level and its personalized prediction sums up the outputs of models assigned to it across all levels. For the top level, FeMAM trains one global model shared by all clients as FedAvg. For every mid-level, it learns multiple models each assigned to a subgroup of clients, as clustered FL. Every bottom-level model is trained for one client only. In the training objective, each model aims to minimize the residual of the additive predictions by the other models assigned to each client. To approximate the arbitrary structure of non-IID across clients, FeMAM introduces more flexibility and adaptivity to FL by incrementally adding new models to the prediction of each client and reassigning another if necessary, automatically optimizing the knowledge-sharing structure. Extensive experiments show that FeMAM surpasses existing clustered FL and personalized FL methods in various non-IID settings. Our code is available at https://github.com/shutong043/FeMAM.

The metaverse has recently gained increasing attention from the public. It builds up a virtual world where we can live as a new role regardless of the role we play in the physical world. However, building and operating this virtual world will generate an extraordinary amount of carbon emissions for computing, communicating, displaying, and so on. This inevitably hinders the realization of carbon neutrality as a priority of our society, adding heavy burden to our earth. In this survey, we first present a green viewpoint of the metaverse by investigating the carbon issues in its three core layers, namely the infrastructure layer, the interaction layer, and the economy layer, and estimate their carbon footprints in the near future. Next, we analyze a range of current and emerging applicable green techniques for the purpose of reducing energy usage and carbon emissions of the metaverse, and discuss their limitations in supporting metaverse workloads. Then, in view of these limitations, we discuss important implications and bring forth several insights and future directions to make each metaverse layer greener. After that, we investigate green solutions from the governance perspective, including both public policies in the physical world and regulation of users in the virtual world, and propose an indicator Carbon Utility (CU) to quantify the service quality brought by an user activity per unit of carbon emissions. Finally, we identify an issue for the metaverse as a whole and summarize three directions: (1) a comprehensive consideration of necessary performance metrics, (2) a comprehensive consideration of involved layers and multiple internal components, and (3) a new assessing, recording, and regulating mechanism on carbon footprints of user activities. Our proposed quantitative indicator CU would be helpful in regulating user activities in the metaverse world., Comment: 24 pages

Lots of real-time applications over Internet of things (IoT)-based status update systems have imperative demands on information freshness, which is usually evaluated by age of information (AoI). Compared to the average AoI and peak AoI (PAoI), violation probabilities and distributions of AoI and PAoI characterize the timeliness in more details. This paper studies the timeliness of the IoT-based multi-source status update system. By modeling the system as a multi-source M/G/1/1 bufferless preemptive queue, general formulas of violation probabilities and probability density functions (p.d.f.s) of AoI and PAoI are derived with a time-domain approach. For the case with negativeexponentially distributed service time, the violation probabilities and p.d.f.s are obtained in closed form. Moreover, the maximal violation probabilities of AoI and PAoI are proposed to characterize the overall timeliness. To improve the overall timeliness under the resource constraint of IoT-device, the arrival rate allocation scheme is used to minimize the maximal violation probabilities. It is proved that the optimal arrival rates can be found by convex optimization algorithms. In addition, it is obtained that the minimum of maximal violation probability of AoI (or PAoI) is achieved only if all violation probabilities of AoI (or PAoI) are equal. Finally, numerical results verify the theoretical analysis and show the effectiveness of the arrival rate allocation scheme.

In Internet of Things (IoT), the decision timeliness of time-sensitive applications is jointly affected by the statistics of update process and decision process. This work considers an update-and-decision system with a Poisson-arrival bufferless queue, where updates are delivered and processed for making decisions with exponential or periodic intervals. We use age-upon-decisions (AuD) to characterize timeliness of updates at decision moments, and the missing probability to specify whether updates are useful for decision-making. Our theoretical analyses 1) present the average AuDs and the missing probabilities for bufferless systems with exponential or deterministic decision intervals under different service time distributions; 2) show that for service scheduling, the deterministic service time achieves a lower average AuD and a smaller missing probability than the uniformly distributed and the negative exponentially distributed service time; 3) prove that the average AuD of periodical decision system is larger than and will eventually drop to that of Poisson decision system along with the increase of decision rate; however, the missing probability in periodical decision system is smaller than that of Poisson decision system. The numerical results and simulations verify the correctness of our analyses, and demonstrate that the bufferless systems outperform the systems applying infinite buffer length.