arXiv subjects

Nicolò Michelusi

Publications and source records attributed to Nicolò Michelusi.

27 records · Page 2Linked to original sources

arXiv2020

Federated Learning with Communication Delay in Edge Networks

Federated learning has received significant attention as a potential solution for distributing machine learning (ML) model training through edge networks. This work addresses an important consideration of federated learning at the network edge: communication delays between the edge nodes and the aggregator. A technique called FedDelAvg (federated delayed averaging) is developed, which generalizes the standard federated averaging algorithm to incorporate a weighting between the current local model and the delayed global model received at each device during the synchronization step. Through theoretical analysis, an upper bound is derived on the global model loss achieved by FedDelAvg, which reveals a strong dependency of learning performance on the values of the weighting and learning rate. Experimental results on a popular ML task indicate significant improvements in terms of convergence speed when optimizing the weighting scheme to account for delays.

cs.LG↗

arXiv2020

Adaptive Millimeter-Wave Communications Exploiting Mobility and Blockage Dynamics

Mobility may degrade the performance of next-generation vehicular networks operating at the millimeter-wave spectrum: frequent loss of alignment and blockages require repeated beam training and handover, thus incurring huge overhead. In this paper, an adaptive and joint design of beam training, data transmission and handover is proposed, that exploits the mobility process of mobile users and the dynamics of blockages to optimally trade-off throughput and power consumption. At each time slot, the serving base station decides to perform either beam training, data communication, or handover when blockage is detected. The problem is cast as a partially observable Markov decision process, and solved via an approximate dynamic programming algorithm based on PERSEUS [2]. Numerical results show that the PERSEUS-based policy performs near-optimally, and achieves a 55% gain in spectral efficiency compared to a baseline scheme with periodic beam training. Inspired by its structure, an adaptive heuristic policy is proposed with low computational complexity and small performance degradation.

cs.IT↗

arXiv2020

Finite rate distributed weight-balancing and average consensus over digraphs

This paper proposes the first distributed algorithm that solves the weight-balancing problem using only finite rate and simplex communications among nodes, compliant with the directed nature of the graph edges. It is proved that the algorithm converges to a weight-balanced solution at sublinear rate. The analysis builds upon a new metric inspired by positional system representations, which characterizes the dynamics of information exchange over the network, and on a novel step-size rule. Building on this result, a novel distributed algorithm is proposed that solves the average consensus problem over digraphs, using, at each timeslot, finite rate simplex communications between adjacent nodes -- some bits for the weight-balancing problem and others for the average consensus. Convergence of the proposed quantized consensus algorithm to the average of the node's unquantized initial values is established, both almost surely and in the moment generating function of the error; and a sublinear convergence rate is proved for sufficiently large step-sizes. Numerical results validate our theoretical findings.

math.OC↗

arXiv2020

Power-Constrained Trajectory Optimization for Wireless UAV Relays with Random Requests

This paper studies the adaptive trajectory design of a rotary-wing UAV serving as a relay between ground nodes dispersed in a circular cell and generating uplink data transmissions randomly according to a Poisson process, and a central base station. We seek to minimize the expected average communication delay to service the data transmission requests, subject to an average power constraint on the mobility of the UAV. The problem is cast as a semi-Markov decision process, and it is shown that the policy exhibits a two-scale structure, which can be efficiently optimized: in the outer decision, upon starting a communication phase, and given its current radius, the UAV selects a target end radius position so as to optimally balance a trade-off between average long-term communication delay and power consumption; in the inner decision, the UAV selects its trajectory between the start radius and the selected end radius, so as to greedily minimize the delay and energy consumption to serve the current request. Numerical evaluations show that, during waiting phases, the UAV circles at some optimal radius at the most energy efficient speed, until a new request is received. Lastly, the expected average communication delay and power consumption of the optimal policy is compared to that of several heuristics, demonstrating a reduction in latency by over 50% and 20%, respectively, compared to static and mobile heuristic schemes.

cs.IT↗

arXiv2019

Trajectory Optimization for Rotary-Wing UAVs in Wireless Networks with Random Requests

This paper studies the trajectory optimization problem in a scenario where a single rotary-wing UAV acts as a relay of data payloads for downlink transmission requests generated randomly by two ground nodes (GNs) in a wireless network. The goal is to optimize the UAV trajectory in order to minimize the expected average communication delay to serve these random requests. It is shown that the problem can be cast as a semi-Markov decision process (SMDP), and the resulting minimization problem is solved via multi-chain policy iteration. The optimality of a two-scale optimization approach is proved: the optimal trajectory in the communication phase greedily minimizes the communication delay of the current request while moving between the current start position and a target end position (inner optimization); the end positions are selected to minimize the expected average long-term delay in the SMDP (outer optimization). Numerical simulations show that the expected average delay is minimized when the UAV moves towards the geometric center of the GNs during phases in which it is not actively servicing transmission requests, and demonstrate significant improvements over sensible heuristics. Finally, it is revealed that the optimal end positions of communication phases become increasingly independent of the data payload, for large data payload values.

cs.IT↗

arXiv2018

Limited Rate Distributed Weight-Balancing and Average Consensus Over Digraphs

Distributed quantized weight-balancing and average consensus over fixed digraphs are considered. A digraph with non-negative weights associated to its edges is weight-balanced if, for each node, the sum of the weights of its out-going edges is equal to that of its incoming edges. This paper proposes and analyzes the first distributed algorithm that solves the weight-balancing problem using only finite rate and simplex communications among nodes (compliant to the directed nature of the graph edges). Asymptotic convergence of the scheme is proved and a convergence rate analysis is provided. Building on this result, a novel distributed algorithm is proposed that solves the average consensus problem over digraphs, using, at each iteration, finite rate simplex communications between adjacent nodes -- some bits for the weight-balancing problem, other for the average consensus. Convergence of the proposed quantized consensus algorithm to the average of the real (i.e., unquantized) agent's initial values is proved, both almost surely and in $r$th mean for all positive integer $r$. Finally, numerical results validate our theoretical findings.

math.OC↗

arXiv2018

Beam Training and Data Transmission Optimization in Millimeter-Wave Vehicular Networks

Future vehicular communication networks call for new solutions to support their capacity demands, by leveraging the potential of the millimeter-wave (mm-wave) spectrum. Mobility, in particular, poses severe challenges in their design, and as such shall be accounted for. A key question in mm-wave vehicular networks is how to optimize the trade-off between directive Data Transmission (DT) and directional Beam Training (BT), which enables it. In this paper, learning tools are investigated to optimize this trade-off. In the proposed scenario, a Base Station (BS) uses BT to establish a mm-wave directive link towards a Mobile User (MU) moving along a road. To control the BT/DT trade-off, a Partially Observable (PO) Markov Decision Process (MDP) is formulated, where the system state corresponds to the position of the MU within the road link. The goal is to maximize the number of bits delivered by the BS to the MU over the communication session, under a power constraint. The resulting optimal policies reveal that adaptive BT/DT procedures significantly outperform common-sense heuristic schemes, and that specific mobility features, such as user position estimates, can be effectively used to enhance the overall system performance and optimize the available system resources.

cs.IT↗

arXiv2016

Queuing models for abstracting interactions in Bacterial communities

Microbial communities play a significant role in bioremediation,plant growth,human and animal digestion,global elemental cycles including the carbon-cycle,and water treatment.They are also posed to be the engines of renewable energy via microbial fuel cells which can reverse the process of electrosynthesis.Microbial communication regulates many virulence mechanisms used by bacteria.Thus,it is of fundamental importance to understand interactions in microbial communities and to develop predictive tools that help control them,in order to aid the design of systems exploiting bacterial capabilities.This position paper explores how abstractions from communications,networking and information theory can play a role in understanding and modeling bacterial interactions.In particular,two forms of interactions in bacterial systems will be examined:electron transfer and quorum sensing.While the diffusion of chemical signals has been heavily studied,electron transfer occurring in living cells and its role in cell-cell interaction is less understood.Recent experimental observations open up new frontiers in the design of microbial systems based on electron transfer,which may coexist with the more well-known interaction strategies based on molecular diffusion.In quorum sensing,the concentration of certain signature chemical compounds emitted by the bacteria is used to estimate the bacterial population size,so as to activate collective behaviors.In this position paper,queuing models for electron transfer are summarized and adapted to provide new models for quorum sensing.These models are stochastic,and thus capture the inherent randomness exhibited by cell colonies in nature.It is shown that queuing models allow the characterization of the state of a single cell as a function of interactions with other cells and the environment,while being amenable to complexity reduction.

cs.ET↗

arXiv2013

Cognitive Access Policies under a Primary ARQ process via Forward-Backward Interference Cancellation

This paper introduces a novel technique for access by a cognitive Secondary User (SU) using best-effort transmission to a spectrum with an incumbent Primary User (PU), which uses Type-I Hybrid ARQ. The technique leverages the primary ARQ protocol to perform Interference Cancellation (IC) at the SU receiver (SUrx). Two IC mechanisms that work in concert are introduced: Forward IC, where SUrx, after decoding the PU message, cancels its interference in the (possible) following PU retransmissions of the same message, to improve the SU throughput; Backward IC, where SUrx performs IC on previous SU transmissions, whose decoding failed due to severe PU interference. Secondary access policies are designed that determine the secondary access probability in each state of the network so as to maximize the average long-term SU throughput by opportunistically leveraging IC, while causing bounded average long-term PU throughput degradation and SU power expenditure. It is proved that the optimal policy prescribes that the SU prioritizes its access in the states where SUrx knows the PU message, thus enabling IC. An algorithm is provided to optimally allocate additional secondary access opportunities in the states where the PU message is unknown. Numerical results are shown to assess the throughput gain provided by the proposed techniques.

cs.IT↗