arXiv subjects

Mohamed Cheriet

Publications and source records attributed to Mohamed Cheriet.

At least 19 recordsLinked to original sources

arXiv2026

Empowering Rural Areas with Multi-radio Microwave Backhaul Supported by Digital Twin for 5G IAB-based FWA

For digital inclusion, high-capacity Internet access should be provided to rural areas to support a range of services and applications. Due to the high operating costs of fiber-optic deployment, Fixed Wireless Access (FWA) is becoming a more attractive internet solution for rural areas. However, 5G FWA is a one-hop solution with limited coverage. A multi-hop solution is needed for wider rural coverage. This work considers a unified solution combining long-haul microwave, 5G Integrated Access and Backhaul (IAB), and FWA to provide a multi-hop network for extended coverage and high network capacity in rural areas. A key challenge for such a network is that energy consumption increases with the number of hops, a problem that has been overlooked in the existing literature. To address this, we propose energy-efficiency microwave backhaul for IAB-based FWA as the Physical Twin (PT). We develop an energy-efficient strategy to optimize radio start-up, serving, sleeping, and wake-up states for microwave backhaul connecting 5G IAB-based FWA serving rural areas. By operating the network at reduced capacity during low utilization, we aim to minimize energy consumption. Then, we present a Digital Twin (DT) of PT to improve its performance. We solve the formulated optimization problem using deep Q-learning in DT and the optimization solver in PT. The simulation results show that our approach satisfies the data rate requirements while reducing energy consumption.

cs.NI↗

arXiv2026

Large Language Model Assisted Intent-Based Satellite-Integrated Access and Backhaul FWA for Rural Areas

Rural areas exhibit low population density and highly variable connectivity needs shaped by both household usage and field operations such as planting, harvesting, and mining. These field activities often occur in isolated locations requiring temporary connectivity, whereas rural households depend on fixed broadband. During intensive outdoor activities, household fixed networks may remain underutilized, resulting in inefficient resource use and unnecessary energy consumption. The coexistence of residential and field-based communication demands creates substantial spatial and temporal fluctuations that the current rural network cannot effectively adapt to. Limited visibility into user mobility, activity patterns, and intent makes it difficult for operators to coordinate temporary and fixed networks. To address these underexplored challenges, we propose an AI driven Intent Aware Satellite Integrated Access and Backhaul (IAB) approach to connect rural areas. In our proposal, a large language model (LLM) translates users' intents into explicit network requirements. Guided by these inferred requirements, we develop a dynamic satellite IAB based Fixed Wireless Access (FWA) network approach that jointly optimizes temporary field connectivity and fixed broadband access to maximize energy efficiency while satisfying the data rate requirement. The formulated optimization problem is solved using a two stage Benders decomposition approach. The simulation results show that our approach significantly reduces energy consumption while maximizing energy efficiency.

cs.NI↗

arXiv2026

FAIR-Pruner: A Flexible Framework for Automatic Layer-Wise Pruning via Tolerance of Difference

Structured pruning is a standard tool for compressing deep neural networks, but its practical performance depends on how sparsity is allocated across layers. We propose FAIR-Pruner, a search-free framework for adaptive layer-wise structured pruning. FAIR-Pruner uses two within-layer rankings: a removal-oriented signal that proposes candidate units and a protection-oriented signal that identifies task-sensitive units. Its core component, Tolerance of Difference (ToD), measures the overlap between the removal prefix and the protected tail, and uses a shared tolerance level to induce non-uniform pruning depths across layers. As a default vision instantiation, FAIR-Pruner combines a Wasserstein-based U-Score for class-conditional unit separability with a Taylor-based R-Score for task-level sensitivity; the same ToD allocation rule can also be paired with alternative removal signals. Theoretically, we analyze ToD through the population R-Score, derive rank-based control of the high-R-Score mass entering the pruning set, and identify an additive exchange condition for same-budget comparison with uniform pruning. Experiments on CIFAR-10, CIFAR-100, SVHN, and ImageNet across VGG, ResNet, DenseNet, ConvNeXt, and DeiT show strong accuracy--compression trade-offs. Prune-only experiments on routed-expert Qwen1.5-MoE-A2.7B-Chat further examine architectural extensibility under matched expert budgets. FAIR-Pruner is released as a pip-installable open-source package.

cs.LG↗

arXiv2026

Architecture-Agnostic Curriculum Learning for Document Understanding: Empirical Evidence from Text-Only and Multimodal

We investigate whether progressive data scheduling -- a curriculum learning strategy that incrementally increases training data exposure (33\%$\rightarrow$67\%$\rightarrow$100\%) -- yields consistent efficiency gains across architecturally distinct document understanding models. By evaluating BERT (text-only, 110M parameters) and LayoutLMv3 (multimodal, 126M parameters) on the FUNSD and CORD benchmarks, we establish that this schedule reduces wall-clock training time by approximately 33\%, commensurate with the reduction from 6.67 to 10.0 effective epoch-equivalents of data. To isolate curriculum effects from compute reduction, we introduce matched-compute baselines (Standard-7) that control for total gradient updates. On the FUNSD dataset, the curriculum significantly outperforms the matched-compute baseline for BERT ($Δ$F1 = +0.023, $p=0.022$, $d_z=3.83$), constituting evidence for a genuine scheduling benefit in capacity-constrained models. In contrast, no analogous benefit is observed for LayoutLMv3 ($p=0.621$), whose multimodal representations provide sufficient inductive bias. On the CORD dataset, all conditions converge to equivalent F1 scores ($\geq$0.947) irrespective of scheduling, indicating a performance ceiling. Schedule ablations comparing progressive, two-phase, reverse, and random pacing confirm that the efficiency gain derives from reduced data volume rather than ordering. Taken together, these findings demonstrate that progressive scheduling is a reliable compute-reduction strategy across model families, with curriculum-specific benefits contingent on the interaction between model capacity and task complexity.

cs.CL↗

arXiv2025

Energy-Efficient Multi-Radio Microwave and IAB-Based Fixed Wireless Access for Rural Areas

Deploying fiber optics as a last-mile solution in rural areas is not economically viable due to low population density. Nevertheless, providing high-speed internet access in these regions is essential to promote digital inclusion. 5G Fixed Wireless Access (5G FWA) has emerged as a promising alternative; however, its one-hop topology limits coverage. To overcome this limitation, a multi-hop architecture is required. This work proposes a unified multi-hop framework that integrates long-haul microwave, Integrated Access and Backhaul (IAB), and FWA to provide wide coverage and high capacity in rural areas. As the number of hops increases, total energy consumption also rises, a challenge often overlooked in existing literature. To address this, we propose an energy-efficient multi-radio microwave and IAB-based FWA framework for rural area connectivity. When the network is underutilized, the proposed approach dynamically operates at reduced capacity to minimize energy consumption. We optimize the off, start-up, serving, deep sleep, and wake-up sates of microwave radios to balance energy use and satisfying data rate requirements. Additionally, we optimize resource block allocation for IAB-based FWA nodes connected to microwave backhaul. The formulated optimization problems aim to minimize the energy consumption of long-haul microwave and multi-hop IAB-based network while satisfying data rate constraints. These problems are solved using dual decomposition and multi-convex programming, supported by dynamic programming. Simulation results demonstrates

cs.NI↗

arXiv2024

HAND: Hierarchical Attention Network for Multi-Scale Handwritten Document Recognition and Layout Analysis

Handwritten document recognition (HDR) is one of the most challenging tasks in the field of computer vision, due to the various writing styles and complex layouts inherent in handwritten texts. Traditionally, this problem has been approached as two separate tasks, handwritten text recognition and layout analysis, and struggled to integrate the two processes effectively. This paper introduces HAND (Hierarchical Attention Network for Multi-Scale Document), a novel end-to-end and segmentation-free architecture for simultaneous text recognition and layout analysis tasks. Our model's key components include an advanced convolutional encoder integrating Gated Depth-wise Separable and Octave Convolutions for robust feature extraction, a Multi-Scale Adaptive Processing (MSAP) framework that dynamically adjusts to document complexity and a hierarchical attention decoder with memory-augmented and sparse attention mechanisms. These components enable our model to scale effectively from single-line to triple-column pages while maintaining computational efficiency. Additionally, HAND adopts curriculum learning across five complexity levels. To improve the recognition accuracy of complex ancient manuscripts, we fine-tune and integrate a Domain-Adaptive Pre-trained mT5 model for post-processing refinement. Extensive evaluations on the READ 2016 dataset demonstrate the superior performance of HAND, achieving up to 59.8% reduction in CER for line-level recognition and 31.2% for page-level recognition compared to state-of-the-art methods. The model also maintains a compact size of 5.60M parameters while establishing new benchmarks in both text recognition and layout analysis. Source code and pre-trained models are available at : https://github.com/MHHamdan/HAND.

cs.CV↗

arXiv2024

HTR-JAND: Handwritten Text Recognition with Joint Attention Network and Knowledge Distillation

Despite significant advances in deep learning, current Handwritten Text Recognition (HTR) systems struggle with the inherent complexity of historical documents, including diverse writing styles, degraded text quality, and computational efficiency requirements across multiple languages and time periods. This paper introduces HTR-JAND (HTR-JAND: Handwritten Text Recognition with Joint Attention Network and Knowledge Distillation), an efficient HTR framework that combines advanced feature extraction with knowledge distillation. Our architecture incorporates three key components: (1) a CNN architecture integrating FullGatedConv2d layers with Squeeze-and-Excitation blocks for adaptive feature extraction, (2) a Combined Attention mechanism fusing Multi-Head Self-Attention with Proxima Attention for robust sequence modeling, and (3) a Knowledge Distillation framework enabling efficient model compression while preserving accuracy through curriculum-based training. The HTR-JAND framework implements a multi-stage training approach combining curriculum learning, synthetic data generation, and multi-task learning for cross-dataset knowledge transfer. We enhance recognition accuracy through context-aware T5 post-processing, particularly effective for historical documents. Comprehensive evaluations demonstrate HTR-JAND's effectiveness, achieving state-of-the-art Character Error Rates (CER) of 1.23\%, 1.02\%, and 2.02\% on IAM, RIMES, and Bentham datasets respectively. Our Student model achieves a 48\% parameter reduction (0.75M versus 1.5M parameters) while maintaining competitive performance through efficient knowledge transfer. Source code and pre-trained models are available at \href{https://github.com/DocumentRecognitionModels/HTR-JAND}{Github}.

cs.CV↗

arXiv2024

Renewable Energy Powered and Open RAN-based Architecture for 5G Fixed Wireless Access Provisioning in Rural Areas

Due to the high costs of optical fiber deployment in Low-Density and Rural Areas (LDRAs), 5G Fixed Wireless Access (5G FWA) recently emerged as an affordable solution. A widely adopted deployment scenario of 5G FWA includes edge cloud that supports computing services and Radio Access Network (RAN) functions. Such edge cloud requires network and energy resources for 5G FWA. This paper proposes renewable energy powered and Open RAN-based architecture for 5G FWA serving LDRAs using three-level closed-loops. Open RAN is a new 5G RAN architecture allowing Open Central Unit and Open Distributed Unit to be distributed in virtualized environment. The first closed-loop distributes radio resources to Open RAN instances and slices at the edge cloud. The second closed-loop allocates radio resources to houses. We design a new energy model that leverages renewable energy. We jointly optimize radio and energy resource allocation in closed-loop 3. We formulate ultra-small and small-time scale optimization problems that link closed-loops to maximize communication utility while minimizing energy costs. We propose reinforcement learning and successive convex approximation to solve the formulated problems. Then, we use solution data and continual learning to improve resource allocation on a large timescale. Our proposal satisfies 97.14% slice delay budget.

cs.NI↗

arXiv2024

Digital Twin Backed Closed-Loops for Energy-Aware and Open RAN-based Fixed Wireless Access Serving Rural Areas

Internet access in rural areas should be improved to support digital inclusion and 5G services. Due to the high deployment costs of fiber optics in these areas, Fixed Wireless Access (FWA) has become a preferable alternative. Additionally, the Open Radio Access Network (O-RAN) can facilitate the interoperability of FWA elements, allowing some FWA functions to be deployed at the edge cloud. However, deploying edge clouds in rural areas can increase network and energy costs. To address these challenges, we propose a closed-loop system assisted by a Digital Twin (DT) to automate energy-aware O-RAN based FWA resource management in rural areas. We consider the FWA and edge cloud as the Physical Twin (PT) and design a closed-loop that distributes radio resources to edge cloud instances for scheduling. We develop another closed-loop for intra-slice resource allocation to houses. We design an energy model that integrates radio resource allocation and formulate ultra-small and small-timescale optimizations for the PT to maximize slice requirement satisfaction while minimizing energy costs. We then design a reinforcement learning approach and successive convex approximation to address the formulated problems. We present a DT that replicates the PT by incorporating solution experiences into future states. The results show that our approach efficiently uses radio and energy resources.

cs.NI↗

arXiv2024

Small Contributions, Small Networks: Efficient Neural Network Pruning Based on Relative Importance

Recent advancements have scaled neural networks to unprecedented sizes, achieving remarkable performance across a wide range of tasks. However, deploying these large-scale models on resource-constrained devices poses significant challenges due to substantial storage and computational requirements. Neural network pruning has emerged as an effective technique to mitigate these limitations by reducing model size and complexity. In this paper, we introduce an intuitive and interpretable pruning method based on activation statistics, rooted in information theory and statistical analysis. Our approach leverages the statistical properties of neuron activations to identify and remove weights with minimal contributions to neuron outputs. Specifically, we build a distribution of weight contributions across the dataset and utilize its parameters to guide the pruning process. Furthermore, we propose a Pruning-aware Training strategy that incorporates an additional regularization term to enhance the effectiveness of our pruning method. Extensive experiments on multiple datasets and network architectures demonstrate that our method consistently outperforms several baseline and state-of-the-art pruning techniques.

cs.LG↗

arXiv2023

Age of Processing-Based Data Offloading for Autonomous Vehicles in Multi-RATs Open RAN

Today, vehicles use smart sensors to collect data from the road environment. This data is often processed onboard of the vehicles, using expensive hardware. Such onboard processing increases the vehicle's cost, quickly drains its battery, and exhausts its computing resources. Therefore, offloading tasks onto the cloud is required. Still, data offloading is challenging due to low latency requirements for safe and reliable vehicle driving decisions. Moreover, age of processing was not considered in prior research dealing with low-latency offloading for autonomous vehicles. This paper proposes an age of processing-based offloading approach for autonomous vehicles using unsupervised machine learning, Multi-Radio Access Technologies (multi-RATs), and Edge Computing in Open Radio Access Network (O-RAN). We design a collaboration space of edge clouds to process data in proximity to autonomous vehicles. To reduce the variation in offloading delay, we propose a new communication planning approach that enables the vehicle to optimally preselect the available RATs such as Wi-Fi, LTE, or 5G to offload tasks to edge clouds when its local resources are insufficient. We formulate an optimization problem for age-based offloading that minimizes elapsed time from generating tasks and receiving computation output. To handle this non-convex problem, we develop a surrogate problem. Then, we use the Lagrangian method to transform the surrogate problem to unconstrained optimization problem and apply the dual decomposition method. The simulation results show that our approach significantly minimizes the age of processing in data offloading with 90.34 % improvement over similar method.

cs.NI↗

arXiv2023

Federated Learning Assisted Deep Q-Learning for Joint Task Offloading and Fronthaul Segment Routing in Open RAN

Offloading computation-intensive tasks to edge clouds has become an efficient way to support resource constraint edge devices. However, task offloading delay is an issue largely due to the networks with limited capacities between edge clouds and edge devices. In this paper, we consider task offloading in Open Radio Access Network (O-RAN), which is a new 5G RAN architecture allowing Open Central Unit (O-CU) to be co-located with Open Distributed Unit (DU) at the edge cloud for low-latency services. O-RAN relies on fronthaul network to connect O-RAN Radio Units (O-RUs) and edge clouds that host O-DUs. Consequently, tasks are offloaded onto the edge clouds via wireless and fronthaul networks \cite{10045045}, which requires routing. Since edge clouds do not have the same available computation resources and tasks' computation deadlines are different, we need a task distribution approach to multiple edge clouds. Prior work has never addressed this joint problem of task offloading, fronthaul routing, and edge computing. To this end, using segment routing, O-RAN intelligent controllers, and multiple edge clouds, we formulate an optimization problem to minimize offloading, fronthaul routing, and computation delays in O-RAN. To determine the solution of this NP-hard problem, we use Deep Q-Learning assisted by federated learning with a reward function that reduces the Cost of Delay (CoD). The simulation results show that our solution maximizes the reward in minimizing CoD.

cs.NI↗

arXiv2023

Evolution of MAC Protocols in the Machine Learning Decade: A Comprehensive Survey

The last decade, (2012 - 2022), saw an unprecedented advance in machine learning (ML) techniques, particularly deep learning (DL). As a result of the proven capabilities of DL, a large amount of work has been presented and studied in almost every field. Since 2012, when the convolution neural networks have been reintroduced in the context of \textit{ImagNet} competition, DL continued to achieve superior performance in many challenging tasks and problems. Wireless communications, in general, and medium access control (MAC) techniques, in particular, were among the fields that were heavily affected by this improvement. MAC protocols play a critical role in defining the performance of wireless communication systems. At the same time, the community lacks a comprehensive survey that collects, analyses, and categorizes the recent work in ML-inspired MAC techniques. In this work, we fill this gap by surveying a long line of work in this era. We solidify the impact of machine learning on wireless MAC protocols. We provide a comprehensive background to the widely adopted MAC techniques, their design issues, and their taxonomy, in connection with the famous application domains. Furthermore, we provide an overview of the ML techniques that have been considered in this context. Finally, we augment our work by proposing some promising future research directions and open research questions that are worth further investigation.

cs.NI↗

arXiv2022

Non-Negative Matrix Factorization with Scale Data Structure Preservation

The model described in this paper belongs to the family of non-negative matrix factorization methods designed for data representation and dimension reduction. In addition to preserving the data positivity property, it aims also to preserve the structure of data during matrix factorization. The idea is to add, to the NMF cost function, a penalty term to impose a scale relationship between the pairwise similarity matrices of the original and transformed data points. The solution of the new model involves deriving a new parametrized update scheme for the coefficient matrix, which makes it possible to improve the quality of reduced data when used for clustering and classification. The proposed clustering algorithm is compared to some existing NMF-based algorithms and to some manifold learning-based algorithms when applied to some real-life datasets. The obtained results show the effectiveness of the proposed algorithm.

cs.LG↗

arXiv2022

Two-level Closed Loops for RAN Slice Resources Management Serving Flying and Ground-based Cars

Flying and ground-based cars require various services such as autonomous driving, remote pilot, infotainment, and remote diagnosis. Each service requires specific Quality of Service (QoS) and network features. Therefore, network slicing can be a solution to fulfill the requirements of various services. Some services, such as infotainment, may have similar requirements to serve flying and ground-based cars. Therefore, some slices can serve both kinds of cars. However, when network slice resource sharing is too aggressive, slices can not meet QoS requirements, where resource under-provisioning causes the violation of QoS, and resource over-provisioning causes resource under-utilization. We propose two closed loops for managing RAN slice resources for cars to address these challenges. First, we present an auction mechanism for allocating Resource Block (RB) to the tenants who provide services to the cars using slices. Second, we design one closed loop that maps slices and services of tenants to virtual Open Distributed Units (vO-DUs) and assigns RB to vO-DUs for management purposes. Third, we design another closed loop for intra-slices RB scheduling to serve cars. Fourth, we present a reward function that interconnects these two closed loops to satisfy the time-varying demands of cars at each slice while meeting QoS requirements in terms of delay. Finally, we design distributed deep reinforcement learning approach to maximize the formulated reward function. The simulation results show that our approach satisfies more than 90% vO-DUs resource constraints and network slice requirements.

cs.NI↗

arXiv2022

PRVNet: A Novel Partially-Regularized Variational Autoencoders for Massive MIMO CSI Feedback

In a multiple-input multiple-output frequency-division duplexing (MIMO-FDD) system, the user equipment (UE) sends the downlink channel state information (CSI) to the base station to report link status. Due to the complexity of MIMO systems, the overhead incurred in sending this information negatively affects the system bandwidth. Although this problem has been widely considered in the literature, prior work generally assumes an ideal feedback channel. In this paper, we introduce PRVNet, a neural network architecture inspired by variational autoencoders (VAE) to compress the CSI matrix before sending it back to the base station under noisy channel conditions. Moreover, we propose a customized loss function that best suits the special characteristics of the problem being addressed. We also introduce an additional regularization hyperparameter for the learning objective, which is crucial for achieving competitive performance. In addition, we provide an efficient way to tune this hyperparameter using KL-annealing. Experimental results show the proposed model outperforms the benchmark models including two deep learning-based models in a noise-free feedback channel assumption. In addition, the proposed model achieves an outstanding performance under different noise levels for additive white Gaussian noise feedback channels.

cs.NI↗

arXiv2022

Leveraging Centric Data Federated Learning Using Blockchain For Integrity Assurance

Machine learning abilities have become a vital component for various solutions across industries, applications, and sectors. Many organizations seek to leverage AI-based solutions across their business services to unlock better efficiency and increase productivity. Problems, however, can arise if there is a lack of quality data for AI-model training, scalability, and maintenance. We propose a data-centric federated learning architecture leveraged by a public blockchain and smart contracts to overcome this significant issue. Our proposed solution provides a virtual public marketplace where developers, data scientists, and AI-engineer can publish their models and collaboratively create and access quality data for training. We enhance data quality and integrity through an incentive mechanism that rewards contributors for data contribution and verification. Those combined with the proposed framework helped increase with only one user simulation the training dataset with an average of 100 input daily and the model accuracy by approximately 4\%.

cs.LG↗

arXiv2022

A Learning Framework for Bandwidth-Efficient Distributed Inference in Wireless IoT

In wireless Internet of things (IoT), the sensors usually have limited bandwidth and power resources. Therefore, in a distributed setup, each sensor should compress and quantize the sensed observations before transmitting them to a fusion center (FC) where a global decision is inferred. Most of the existing compression techniques and entropy quantizers consider only the reconstruction fidelity as a metric, which means they decouple the compression from the sensing goal. In this work, we argue that data compression mechanisms and entropy quantizers should be co-designed with the sensing goal, specifically for machine-consumed data. To this end, we propose a novel deep learning-based framework for compressing and quantizing the observations of correlated sensors. Instead of maximizing the reconstruction fidelity, our objective is to compress the sensor observations in a way that maximizes the accuracy of the inferred decision (i.e., sensing goal) at the FC. Unlike prior work, we do not impose any assumptions about the observations distribution which emphasizes the wide applicability of our framework. We also propose a novel loss function that keeps the model focused on learning complementary features at each sensor. The results show the superior performance of our framework compared to other benchmark models.

eess.SP↗