Due to the large size of the training data, distributed learning approaches such as federated learning have gained attention recently. However, the convergence rate of distributed learning suffers from heterogeneous worker performance.In this paper, we consider an incentive mechanism for workers to mitigate the delays in completion of each batch. We analytically obtained equilibrium solution of a Stackelberg game. Our numerical results indicate that with a limited budget, the model owner should judiciously decide on the number of workers due to trade off between the diversity provided by the number of workers and the latency of completing the training.Y. Sarikaya (ysarikaya@sabanciuniv.edu) and O. Ercetin (oercetin@sabanciuniv.edu) are with the
Abstract-We consider the problem of cross-layer resource allocation in time-varying cellular wireless networks, and incorporate information theoretic secrecy as a Quality of Service constraint. Specifically, each node in the network injects two types of traffic, private and open, at rates chosen in order to maximize a global utility function, subject to network stability and secrecy constraints. The secrecy constraint enforces an arbitrarily low mutual information leakage from the source to every node in the network, except for the sink node. We first obtain the achievable rate region for the problem for single and multi-user systems assuming that the nodes have full CSI of their neighbors. Then, we provide a joint flow control, scheduling and private encoding scheme, which does not rely on the knowledge of the prior distribution of the gain of any channel. We prove that our scheme achieves a utility, arbitrarily close to the maximum achievable utility. Numerical experiments are performed to verify the analytical results, and to show the efficacy of the dynamic control algorithm.
It is well known that Max-Weight scheduling provides queue stability whenever this is possible. However, MaxWeight scheduling requires the complete channel state information (CSI) to make the best transmission decision at every time slot. The common assumption in this line of research assumes that the network controller has full CSI at every decision time without taking into account the overhead associated with channel probing. In practice, however, acquiring CSI is not cost-free and requires certain amount of resources. In this work, we design a Scheduling and Dynamic Feedback algorithm, named SDF, by considering the overhead of obtaining the channel state information. We first establish a bound on the achievable rate region of SDF algorithm by proving that SDF supports 1 + fraction of of the full rate region (the rate region when all users are probed) where only depends on the expected number of users which are not probed. Then, for homogenous channel, we show that when the number of users in the network is greater than 3, > 0, i.e., we guarantee to expand the rate region. We also demonstrate numerically in a realistic simulation setting that this rate region can be achieved by probing only less than 50% of all channels in a CDMA based cellular network utilizing high data rate protocol under normal channel conditions.
We consider the problem of resource allocation and control of multihop networks in which multiple source-destination pairs communicate confidential messages, to be kept confidential from the intermediate nodes. We pose the problem as that of network utility maximization, into which confidentiality is incorporated as an additional quality of service constraint. We develop a simple, and yet provably optimal dynamic control algorithm that combines flow control, routing and end-to-end secrecy-encoding. In order to achieve confidentiality, our scheme exploits multipath diversity and temporal diversity due to channel variability. Our end-to-end dynamic encoding scheme encodes confidential messages across multiple packets, to be combined at the ultimate destination for recovery. We first develop an optimal dynamic policy for the case in which the number of blocks across which secrecy encoding is performed is asymptotically large. Next, we consider encoding across a finite number of packets, which eliminates the possibility of achieving perfect secrecy. For this case, we develop a dynamic policy to choose the encoding rates for each message, based on the instantaneous channel state information, queue states and secrecy outage requirements. By numerical analysis, we observe that the proposed scheme approaches the optimal rates asymptotically with increasing block size. Finally, we address the consequences of practical implementation issues such as infrequent queue updates and de-centralized scheduling. We demonstrate the efficacy of our policies by numerical studies under various network conditions. Index Terms-.1 It can be proved that this scheme achieves the secrecy capacity of the diamond network, i.e., there exists no other scheme with which one can achieve a rate higher than 1 bit/channel use with perfect secrecy from the relay nodes. In general in a network with parallel and independent paths, th path with capacity identical to , the secrecy capacity is identical to , hence the sum rate minus the rate along the best relay.1063-6692
We consider the problem of resource allocation in a wireless cellular network, in which nodes have both open and private information to be transmitted to the base station over block fading uplink channels. We develop a cross-layer solution, based on hybrid ARQ transmission with incremental redundancy. We provide a scheme that combines power control, flow control, and scheduling in order to maximize a global utility function, subject to the stability of the data queues, an average power constraint, and a constraint on the privacy outage probability. Our scheme is based on the assumption that each node has an estimate of its uplink channel gain at each block, while only the distribution of the cross channel gains is available. We prove that our scheme achieves a utility, arbitrarily close to the maximum achievable utility given the available channel state information.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.