arXiv subjects

Sameh Sorour

Publications and source records attributed to Sameh Sorour.

At least 19 recordsLinked to original sources

arXiv2025

Spatiotemporal Analysis of Parallelized Computing at the Extreme Edge

Extreme Edge Computing (EEC) pushes computing even closer to end users than traditional Multi-access Edge Computing (MEC), harnessing the idle resources of Extreme Edge Devices (EEDs) to enable low-latency, distributed processing. However, EEC faces key challenges, including spatial randomness in device distribution, limited EED computational power necessitating parallel task execution, vulnerability to failure, and temporal randomness due to variability in wireless communication and execution times. These challenges highlight the need for a rigorous analytical framework to evaluate EEC performance. We present the first spatiotemporal mathematical model for EEC over large-scale millimeter-wave networks. Utilizing stochastic geometry and an Absorbing Continuous-Time Markov Chain (ACTMC), the framework captures the complex interaction between communication and computation performance, including their temporal overlap during parallel execution. We evaluate two key metrics: average task response delay and task completion probability. Together, they provide a holistic view of latency and reliability. The analysis considers fundamental offloading strategies, including randomized and location-aware schemes, while accounting for EED failures. Results show that there exists an optimal task segmentation that minimizes delay. Under limited EED availability, we investigate a bias-based EEC and MEC collaboration that offloads excess demand to MEC resources, effectively reducing congestion and improving system responsiveness.

cs.PF↗

arXiv2021

Motivating Learners in Multi-Orchestrator Mobile Edge Learning: A Stackelberg Game Approach

Mobile Edge Learning (MEL) is a learning paradigm that enables distributed training of Machine Learning models over heterogeneous edge devices (e.g., IoT devices). Multi-orchestrator MEL refers to the coexistence of multiple learning tasks with different datasets, each of which being governed by an orchestrator to facilitate the distributed training process. In MEL, the training performance deteriorates without the availability of sufficient training data or computing resources. Therefore, it is crucial to motivate edge devices to become learners and offer their computing resources, and either offer their private data or receive the needed data from the orchestrator and participate in the training process of a learning task. In this work, we propose an incentive mechanism, where we formulate the orchestrators-learners interactions as a 2-round Stackelberg game to motivate the participation of the learners. In the first round, the learners decide which learning task to get engaged in, and then in the second round, the training parameters and the amount of data for training in case of participation such that their utility is maximized. We then study the training round analytically and derive the learners' optimal strategy. Finally, numerical experiments have been conducted to evaluate the performance of the proposed incentive mechanism.

cs.NI↗

arXiv2021

Energy-Efficient Multi-Orchestrator Mobile Edge Learning

Mobile Edge Learning (MEL) is a collaborative learning paradigm that features distributed training of Machine Learning (ML) models over edge devices (e.g., IoT devices). In MEL, possible coexistence of multiple learning tasks with different datasets may arise. The heterogeneity in edge devices' capabilities will require the joint optimization of the learners-orchestrator association and task allocation. To this end, we aim to develop an energy-efficient framework for learners-orchestrator association and learning task allocation, in which each orchestrator gets associated with a group of learners with the same learning task based on their communication channel qualities and computational resources, and allocate the tasks accordingly. Therein, a multi objective optimization problem is formulated to minimize the total energy consumption and maximize the learning tasks' accuracy. However, solving such optimization problem requires centralization and the presence of the whole environment information at a single entity, which becomes impractical in large-scale systems. To reduce the solution complexity and to enable solution decentralization, we propose lightweight heuristic algorithms that can achieve near-optimal performance and facilitate the trade-offs between energy consumption, accuracy, and solution complexity. Simulation results show that the proposed approaches reduce the energy consumption significantly while executing multiple learning tasks compared to recent state-of-the-art methods.

cs.NI↗

arXiv2021

Jointly Optimizing Dataset Size and Local Updates in Heterogeneous Mobile Edge Learning

This paper proposes to maximize the accuracy of a distributed machine learning (ML) model trained on learners connected via the resource-constrained wireless edge. We jointly optimize the number of local/global updates and the task size allocation to minimize the loss while taking into account heterogeneous communication and computation capabilities of each learner. By leveraging existing bounds on the difference between the training loss at any given iteration and the theoretically optimal loss, we derive an expression for the objective function in terms of the number of local updates. The resulting convex program is solved to obtain the optimal number of local updates which is used to obtain the total updates and batch sizes for each learner. The merits of the proposed solution, which is heterogeneity aware (HA), are exhibited by comparing its performance to the heterogeneity unaware (HU) approach.

eess.SP↗

arXiv2020

Task Allocation for Asynchronous Mobile Edge Learning with Delay and Energy Constraints

This paper extends the paradigm of "mobile edge learning (MEL)" by designing an optimal task allocation scheme for training a machine learning model in an asynchronous manner across mutiple edge nodes or learners connected via a resource-constrained wireless edge network. The optimization is done such that the portion of the task allotted to each learner is completed within a given global delay constraint and a local maximum energy consumption limit. The time and energy consumed are related directly to the heterogeneous communication and computational capabilities of the learners; i.e. the proposed model is heterogeneity aware (HA). Because the resulting optimization is an NP-hard quadratically-constrained integer linear program (QCILP), a two-step suggest-and-improve (SAI) solution is proposed based on using the solution of the relaxed synchronous problem to obtain the solution to the asynchronous problem. The proposed HA asynchronous (HA-Asyn) approach is compared against the HA synchronous (HA-Sync) scheme and the heterogeneity unaware (HU) equal batch allocation scheme. Results from a system of 20 learners tested for various completion time and energy consumption constraints show that the proposed HA-Asyn method works better than the HU synchronous/asynchronous (HU-Sync/Asyn) approach and can provide gains of up-to 25\% compared to the HA-Sync scheme.

cs.LG↗

arXiv2020

Adaptive Task Allocation for Asynchronous Federated and Parallelized Mobile Edge Learning

This paper proposes a scheme to efficiently execute distributed learning tasks in an asynchronous manner while minimizing the gradient staleness on wireless edge nodes with heterogeneous computing and communication capacities. The approach considered in this paper ensures that all devices work for a certain duration that covers the time for data/model distribution, learning iterations, model collection and global aggregation. The resulting problem is an integer non-convex program with quadratic equality constraints as well as linear equality and inequality constraints. Because the problem is NP-hard, we relax the integer constraints in order to solve it efficiently with available solvers. Analytical bounds are derived using the KKT conditions and Lagrangian analysis in conjunction with the suggest-and-improve approach. Results show that our approach reduces the gradient staleness and can offer better accuracy than the synchronous scheme and the asynchronous scheme with equal task allocation.

cs.DC↗

arXiv2020

Returning the Favor: What Wireless Networking Can Offer to AI and Edge Learning

Machine learning (ML) and artificial intelligence (AI) have recently made a significant impact on improving the operations of wireless networks and establishing intelligence at the edge. In return, rare efforts were made to explore how adapting, optimizing, and arranging wireless networks can contribute to implementing ML/AI at the edge. This article aims to address this void by setting a vision on how wireless networking researchers can leverage their expertise to return the favor to edge learning. It will review the enabling technologies, summarize the inaugural works on this path, and shed light on different directions to establish a comprehensive framework for mobile edge learning (MEL).

cs.DC↗

arXiv2019

Adaptive Task Allocation for Mobile Edge Learning

This paper aims to establish a new optimization paradigm for implementing realistic distributed learning algorithms, with performance guarantees, on wireless edge nodes with heterogeneous computing and communication capacities. We will refer to this new paradigm as `Mobile Edge Learning (MEL)'. The problem of dynamic task allocation for MEL is considered in this paper with the aim to maximize the learning accuracy, while guaranteeing that the total times of data distribution/aggregation over heterogeneous channels, and local computing iterations at the heterogeneous nodes, are bounded by a preset duration. The problem is first formulated as a quadratically-constrained integer linear problem. Being an NP-hard problem, the paper relaxes it into a non-convex problem over real variables. We thus proposed two solutions based on deriving analytical upper bounds of the optimal solution of this relaxed problem using Lagrangian analysis and KKT conditions, and the use of suggest-and-improve starting from equal batch allocation, respectively. The merits of these proposed solutions are exhibited by comparing their performances to both numerical approaches and the equal task allocation approach.

cs.DC↗

arXiv2018

Throughput Maximization in Cloud-Radio Access Networks using Rate-Aware Network Coding

One of the most promising techniques for network-wide interference management necessitates a redesign of the network architecture known as cloud radio access network (CRAN). The cloud is responsible for coordinating multiple Remote Radio Heads (RRHs) and scheduling users to their radio resources blocks (RRBs). The transmit frame of each RRH consists of several orthogonal RRBs each maintained at a certain power level (PL). While previous works considered a vanilla version in which each RRB can serve a single user, this paper proposes mixing the flows of multiple users using instantly decodable network coding (IDNC). As such, the total throughput is maximized. The joint user scheduling and power adaptation problem is solved by designing, for each RRB, a subgraph in which each vertex represents potential user-RRH associations, encoded files, transmission rates, and PLs for one specific RRB. It is shown that the original problem is equivalent to a maximum-weight clique problem over the union of all subgraphs, called herein the CRAN-IDNCgraph. Extensive simulation results are provided to attest the effectiveness of the proposed solution against state of the art algorithms. In particular, the presented simulation results reveal that the method achieves substantial performance gains for all system configurations which collaborates the theoretical findings.

cs.IT↗

arXiv2018

Deep Learning for IoT Big Data and Streaming Analytics: A Survey

In the era of the Internet of Things (IoT), an enormous amount of sensing devices collect and/or generate various sensory data over time for a wide range of fields and applications. Based on the nature of the application, these devices will result in big or fast/real-time data streams. Applying analytics over such data streams to discover new information, predict future insights, and make control decisions is a crucial process that makes IoT a worthy paradigm for businesses and a quality-of-life improving technology. In this paper, we provide a thorough overview on using a class of advanced machine learning techniques, namely Deep Learning (DL), to facilitate the analytics and learning in the IoT domain. We start by articulating IoT data characteristics and identifying two major treatments for IoT data from a machine learning perspective, namely IoT big data analytics and IoT streaming data analytics. We also discuss why DL is a promising approach to achieve the desired analytics in these types of data and applications. The potential of using emerging DL techniques for IoT data analytics are then discussed, and its promises and challenges are introduced. We present a comprehensive background on different DL architectures and algorithms. We also analyze and summarize major reported research attempts that leveraged DL in the IoT domain. The smart IoT devices that have incorporated DL in their intelligence background are also discussed. DL implementation approaches on the fog and cloud centers in support of IoT applications are also surveyed. Finally, we shed light on some challenges and potential directions for future research. At the end of each section, we highlight the lessons learned based on our experiments and review of the recent literature.

cs.NI↗

arXiv2018

Multi-Class Management with Sub-Class Service for Autonomous Electric Mobility On-Demand Systems

Despite the significant advances in vehicle automation and electrification, the next-decade aspirations for massive deployments of autonomous electric mobility on demand (AEMoD) services are still threatened by two major bottlenecks, namely the computational and charging delays. This paper proposes a solution for these two challenges by suggesting the use of fog computing for AEMoD systems, and developing an optimized charging scheme for its vehicles with and multi-class dispatching scheme for the customers. A queuing model representing the proposed multi-class management scheme with sub-class service is first introduced. The stability conditions of the system in a given city zone are then derived. Decisions on the proportions of each class vehicles to partially/fully charge, or directly serve customers of possible sub-classes are then optimized in order to minimize the maximum response time of the system. Results show the merits of our optimized model compared to a previously proposed scheme and other non-optimized policies.

eess.SP↗

arXiv2018

Throughput Maximization in Cloud Radio Access Networks using Network Coding

This paper is interested in maximizing the total throughput of cloud radio access networks (CRANs) in which multiple radio remote heads (RRHs) are connected to a central computing unit known as the cloud. The transmit frame of each RRH consists of multiple radio resources blocks (RRBs), and the cloud is responsible for synchronizing these RRBS and scheduling them to users. Unlike previous works that consider allocating each RRB to only a single user at each time instance, this paper proposes to mix the flows of multiple users in each RRB using instantly decodable network coding (IDNC). The proposed scheme is thus designed to jointly schedule the users to different RRBs, choose the encoded file sent in each of them, and the rate at which each of them is transmitted. Hence, the paper maximizes the throughput which is defined as the number of correctly received bits. To jointly fulfill this objective, we design a graph in which each vertex represents a possible user-RRB association, encoded file, and transmission rate. By appropriately choosing the weights of vertices, the scheduling problem is shown to be equivalent to a maximum weight clique problem over the newly introduced graph. Simulation results illustrate the significant gains of the proposed scheme compared to classical coding and uncoded solutions.

cs.IT↗

arXiv2018

Optimal Vehicle Dimensioning for Multi-Class Autonomous Electric Mobility On-Demand Systems

Autonomous electric mobility on demand (AEMoD) has recently emerged as a cyber-physical system aiming to bring automation, electrification, and on-demand services for the future private transportation market. The expected massive demand for such services and its resulting insufficient charging time/resources prohibit the use of centralized management and full vehicle charging. A fog-based multi-class solution for these challenges was recently suggested, by enabling per-zone management and partial charging for different classes of AEMoD vehicles. This paper focuses on finding the optimal vehicle dimensioning for each zone of these systems in order to guarantee a bounded response time of its vehicles. Using a queuing model representing the multi-class charging and dispatching processes, we first derive the stability conditions and the number of system classes to guarantee the response time bound. Decisions on the proportions of each class vehicles to partially/fully charge, or directly serve customers are then optimized so as to minimize the vehicles in-flow to any given zone. Excess waiting times of customers in rare critical events, such as limited charging resources and/or limited vehicles availabilities, are also investigated. Results show the merits of our proposed model compared to other schemes and in usual and critical scenarios.

eess.SY↗

arXiv2017

VANETs Meet Autonomous Vehicles: A Multimodal 3D Environment Learning Approach

In this paper, we design a multimodal framework for object detection, recognition and mapping based on the fusion of stereo camera frames, point cloud Velodyne Lidar scans, and Vehicle-to-Vehicle (V2V) Basic Safety Messages (BSMs) exchanged using Dedicated Short Range Communication (DSRC). We merge the key features of rich texture descriptions of objects from 2D images, depth and distance between objects provided by 3D point cloud and awareness of hidden vehicles from BSMs' 3D information. We present a joint pixel to point cloud and pixel to V2V correspondences of objects in frames from the Kitti Vision Benchmark Suite by using a semi-supervised manifold alignment approach to achieve camera-Lidar and camera-V2V mapping of their recognized objects that have the same underlying manifold.

cs.CV↗

arXiv2017

A Multi-Class Dispatching and Charging Scheme for Autonomous Electric Mobility On-Demand

Despite the significant advances in vehicle automation and electrification, the next-decade aspirations for massive deployments of autonomous electric mobility on demand (AEMoD) services are still threatened by two major bottlenecks, namely the computational and charging delays. This paper proposes a solution for these two challenges by suggesting the use of fog computing for AEMoD systems, and developing an optimized multi-class charging and dispatching scheme for its vehicles. A queuing model representing the proposed multi-class charging and dispatching scheme is first introduced. The stability conditions of this model and the number of classes that fit the charging capabilities of any given city zone are then derived. Decisions on the proportions of each class vehicles to partially/fully charge, or directly serve customers are then optimized using a stochastic linear program that minimizes the maximum response time of the system. Results show the merits of our proposed model and optimized decision scheme compared to both the always-charge and the equal split schemes.

eess.SY↗

arXiv2017

Data Dissemination using Instantly Decodable Binary Codes in Fog-Radio Access Networks

Consider a device-to-device (D2D) fog-radio access network wherein a set of devices are required to store a set of files. Each device is connected to a subset of the cloud data centers and thus possesses a subset of the data. This paper investigates the problem of disseminating all files among the devices while reducing the total time of communication, i.e., the completion time, using instantly decodable network coding (IDNC). While previous studies on the use of IDNC in D2D systems assume a fully connected communication network, this paper tackles the more realistic scenario of a partially connected network in which devices can only target devices in their transmission range. The paper first formulates the optimal joint optimization of selecting the transmitting device(s) and the file combination(s) and exhibits its intractability. The completion time is approximated using the celebrated decoding delay approach by deriving the relationship between the quantities in a partially connected network. The paper introduces the cooperation graph and demonstrates that the relaxed problem is equivalent to a maximum weight clique problem over the newly designed graph wherein the weights are obtained by solving a similar problem on the local IDNC graphs. Extensive simulations reveal that the proposed solution provides noticeable performance enhancement and outperforms previously proposed IDNC-based schemes.

cs.IT↗

arXiv2016

Handover Management in 5G and Beyond: A Topology Aware Skipping Approach

Network densification is found to be a potential solution to meet 5G capacity standards. Network densification offers more capacity by shrinking base stations' (BSs) footprints, thus reduces the number of users served by each BS. However, the gains in the capacity are achieved at the expense of increased handover (HO) rates. Hence, HO rate is a key performance limiting factor that should be carefully considered in densification planning. This paper sheds light on the HO problem that appears in dense 5G networks and proposes an effective solution via topology aware HO skipping. Different skipping techniques are considered and compared with the conventional best connected scheme. To this end, the effectiveness of the proposed schemes is validated by studying the average user rate in the downlink single-tier and two-tier cellular networks, which are modeled using Poisson point process and Poisson cluster process, respectively. The proposed skipping schemes show up to 47% gains in the average throughput that would maximize the benefit of network densification.

cs.NI↗

arXiv2016

Velocity-Aware Handover Management in Two-Tier Cellular Networks

While network densification is considered an important solution to cater the ever-increasing capacity demand, its effect on the handover (HO) rate is overlooked. In dense 5G networks, HO delays may neutralize or even negate the gains offered by network densification. Hence, user mobility imposes a nontrivial challenge to harvest capacity gains via network densification. In this paper, we propose a velocity-aware HO management scheme for two-tier downlink cellular network to mitigate the HO effect on the foreseen densification throughput gains. The proposed HO scheme sacrifices the best BS connectivity, by skipping HO to some BSs along the user's trajectory, to maintain longer connection durations and reduce HO rates. Furthermore, the proposed scheme enables cooperative BS service and strongest interference cancellation to compensate for skipping the best connectivity. To this end, we consider different HO skipping scenarios and develop a velocity-aware mathematical model, via stochastic geometry, to quantify the performance of the proposed HO scheme in terms of the coverage probability and user throughput. The results highlight the HO rate problem in dense cellular environments and show the importance of the proposed HO schemes. Finally, the value of BS cooperation along with handover skipping is quantified for different user mobility profiles.

cs.NI↗