Amr Mohamed

Abdelaziz Bounhar, Hadi Abdine, Evan Dufraisse et al.

Large language models (LLMs) trained for step-by-step reasoning often become excessively verbose, raising inference cost. Standard Reinforcement Learning with Verifiable Rewards (RLVR) pipelines filter out ``easy'' problems for training efficiency, leaving the model to train primarily on harder problems that require longer reasoning chains. This skews the output length distribution upward, resulting in a \textbf{model that conflates ``thinking longer'' with ``thinking better''}. In this work, we show that retaining and modestly up-weighting moderately easy problems acts as an implicit length regularizer. Exposing the model to solvable short-chain tasks constrains its output distribution and prevents runaway verbosity. The result is \textbf{\emph{emergent brevity for free}}: the model learns to solve harder problems without inflating the output length, \textbf{ despite the absence of any explicit length penalization}. RLVR experiments using this approach on \textit{Qwen3-4B-Thinking-2507} (with a 16k token limit) achieve baseline pass@1 AIME25 accuracy while generating solutions that are, on average, nearly twice as short. The code is available at \href{https://github.com/MBZUAI-Paris/Frugal-AI}{GitHub}, with datasets and models on \href{https://huggingface.co/collections/MBZUAI-Paris/k2-think-mini-68dcfa8b114686a4bd3dc2bc}{Hugging Face}.

Emna Baccour, Aiman Erbad, Amr Mohamed et al.

Although Deep Neural Networks (DNN) have become the backbone technology of several ubiquitous applications, their deployment in resource-constrained machines, e.g., Internet of Things (IoT) devices, is still challenging. To satisfy the resource requirements of such a paradigm, collaborative deep inference with IoT synergy was introduced. However, the distribution of DNN networks suffers from severe data leakage. Various threats have been presented, including black-box attacks, where malicious participants can recover arbitrary inputs fed into their devices. Although many countermeasures were designed to achieve privacy-preserving DNN, most of them result in additional computation and lower accuracy. In this paper, we present an approach that targets the security of collaborative deep inference via re-thinking the distribution strategy, without sacrificing the model performance. Particularly, we examine different DNN partitions that make the model susceptible to black-box threats and we derive the amount of data that should be allocated per device to hide proprieties of the original input. We formulate this methodology, as an optimization, where we establish a trade-off between the latency of co-inference and the privacy-level of data. Next, to relax the optimal solution, we shape our approach as a Reinforcement Learning (RL) design that supports heterogeneous devices as well as multiple DNNs/datasets.

Noor Khial, Naram Mhaisen, Loay Ismail et al.

Unmanned Aerial Vehicles (UAVs) equipped with radar sensors are deployed for target search missions in diverse environments, where targets exhibit characteristic signatures (e.g., respiration micro-motion in human search) detectable through occlusions. A fundamental challenge arises from shifts in radar statistics as the UAV moves through a dynamic and potentially non-stationary environment, rendering any fixed signal-processing strategy suboptimal; yet perception and adaptation must run onboard a resource-constrained aerial node in real time. Since no single detector performs well across all conditions, we adopt a multi-policy paradigm and formulate UAV target search as an online policy selection problem over a library of specialized detectors, with performance measured by regret, the cumulative loss gap relative to the best policy in each scene. The setting couples in-scene stochastic noise with inter-scene shifts. Whereas prior methods capture only one regime, we account for both through the Stochastically Extended Adversary (SEA) framework, without requiring oracle knowledge of scene dynamics. Because adaptation must run at the UAV, we instantiate SEA through \textsc{SEArch}, a lightweight optimistic Follow the Regularized Leader (OFTRL) selector with an adaptive learning rate, achieving regret $O(\barσ_T \sqrt{T} + \sqrt{J})$, where $\barσ_T$ captures radar measurement noise and $J$ is the number of scene transitions over the mission horizon $T$. To enable rapid adaptation under frequent scene changes, we further introduce \textsc{W-SEArch}, a windowed variant that restarts every $w$ rounds and achieves regret $O(\barσ_I \sqrt{w})$ under at most one transition per window. Experiments show up to 30\% regret reduction compared to non-adaptive baselines across a range of non-stationary settings.

Guokan Shang, Hadi Abdine, Yousef Khoubrane et al.

We introduce Atlas-Chat, the first-ever collection of LLMs specifically developed for dialectal Arabic. Focusing on Moroccan Arabic, also known as Darija, we construct our instruction dataset by consolidating existing Darija language resources, creating novel datasets both manually and synthetically, and translating English instructions with stringent quality control. Atlas-Chat-2B, 9B, and 27B models, fine-tuned on the dataset, exhibit superior ability in following Darija instructions and performing standard NLP tasks. Notably, our models outperform both state-of-the-art and Arabic-specialized LLMs like LLaMa, Jais, and AceGPT, e.g., our 9B model gains a 13% performance boost over a larger 13B model on DarijaMMLU, in our newly introduced evaluation suite for Darija covering both discriminative and generative tasks. Furthermore, we perform an experimental analysis of various fine-tuning strategies and base model choices to determine optimal configurations. All our resources are publicly accessible, and we believe our work offers comprehensive design methodologies of instruction-tuning for low-resource languages, which are often neglected in favor of data-rich languages by contemporary LLMs.

Emna Baccour, Mhd Saria Allahham, Aiman Erbad et al.

The vision of the upcoming 6G technologies, characterized by ultra-dense network, low latency, and fast data rate is to support Pervasive AI (PAI) using zero-touch solutions enabling self-X (e.g., self-configuration, self-monitoring, and self-healing) services. However, the research on 6G is still in its infancy, and only the first steps have been taken to conceptualize its design, investigate its implementation, and plan for use cases. Toward this end, academia and industry communities have gradually shifted from theoretical studies of AI distribution to real-world deployment and standardization. Still, designing an end-to-end framework that systematizes the AI distribution by allowing easier access to the service using a third-party application assisted by a zero-touch service provisioning has not been well explored. In this context, we introduce a novel platform architecture to deploy a zero-touch PAI-as-a-Service (PAIaaS) in 6G networks supported by a blockchain-based smart system. This platform aims to standardize the pervasive AI at all levels of the architecture and unify the interfaces in order to facilitate the service deployment across application and infrastructure domains, relieve the users worries about cost, security, and resource allocation, and at the same time, respect the 6G stringent performance requirements. As a proof of concept, we present a Federated Learning-as-a-service use case where we evaluate the ability of our proposed system to self-optimize and self-adapt to the dynamics of 6G networks in addition to minimizing the users' perceived costs.

Emna Baccour, Aiman Erbad, Amr Mohamed et al.

The metaverse, envisioned as the next digital frontier for avatar-based virtual interaction, involves high-performance models. In this dynamic environment, users' tasks frequently shift, requiring fast model personalization despite limited data. This evolution consumes extensive resources and requires vast data volumes. To address this, meta-learning emerges as an invaluable tool for metaverse users, with federated meta-learning (FML), offering even more tailored solutions owing to its adaptive capabilities. However, the metaverse is characterized by users heterogeneity with diverse data structures, varied tasks, and uneven sample sizes, potentially undermining global training outcomes due to statistical difference. Given this, an urgent need arises for smart coalition formation that accounts for these disparities. This paper introduces a dual game-theoretic framework for metaverse services involving meta-learners as workers to manage FML. A blockchain-based cooperative coalition formation game is crafted, grounded on a reputation metric, user similarity, and incentives. We also introduce a novel reputation system based on users' historical contributions and potential contributions to present tasks, leveraging correlations between past and new tasks. Finally, a Stackelberg game-based incentive mechanism is presented to attract reliable workers to participate in meta-learning, minimizing users' energy costs, increasing payoffs, boosting FML efficacy, and improving metaverse utility. Results show that our dual game framework outperforms best-effort, random, and non-uniform clustering schemes - improving training performance by up to 10%, cutting completion times by as much as 30%, enhancing metaverse utility by more than 25%, and offering up to 5% boost in training efficiency over non-blockchain systems, effectively countering misbehaving users.

Abdulrahman Soliman, Amr Mohamed, Elias Yaacoub et al.

Telemedicine applications have recently received substantial potential and interest, especially after the COVID-19 pandemic. Remote experience will help people get their complex surgery done or transfer knowledge to local surgeons, without the need to travel abroad. Even with breakthrough improvements in internet speeds, the delay in video streaming is still a hurdle in telemedicine applications. This imposes using image compression and region of interest (ROI) techniques to reduce the data size and transmission needs. This paper proposes a Deep Reinforcement Learning (DRL) model that intelligently adapts the ROI size and non-ROI quality depending on the estimated throughput. The delay and structural similarity index measure (SSIM) comparison are used to assess the DRL model. The comparison findings and the practical application reveal that DRL is capable of reducing the delay by 13% and keeping the overall quality in an acceptable range. Since the latency has been significantly reduced, these findings are a valuable enhancement to telemedicine applications.

Amr Mohamed, Yang Zhang, Michalis Vazirgiannis et al.

Diffusion large language models (dLLMs) offer a promising alternative to autoregressive models, but their practical utility is severely hampered by slow, iterative sampling. We present SchED, a training-free, model-agnostic early-exit algorithm that aggregates full-span logit margins and halts decoding once a smooth, progress-dependent confidence threshold is met. We evaluated SchED on two dLLM families (Dream and LLaDA), in base and instruction-tuned variants across ten benchmarks spanning downstream tasks including multiple-choice question answering (MCQ), math, long-form QA/summarization, and translation. SchED delivers large, stable accelerations: on instruction-tuned models, it achieves $3.8$-$4.0\times$ speedups while retaining $99.8$-$100\%$ of the baseline score on average. On base models, SchED yields consistent speedup gains with $99.1$-$100\%$ performance retention, with up to $2.34\times$ under more aggressive settings. Using a conservative speed metric that heavily penalizes quality loss (QPS, $γ{=}4$), we show that SchED is robust and clearly outperforms prior confidence-based early-exit methods, which break down on long-form generation. An entropy analysis of the model's token predictions reveals that instruction tuning speeds up the decay of predictive entropy. By turning genuine confidence stabilization into computational savings, SchED makes dLLM decoding substantially more efficient.

Mingmeng Geng, Amr Mohamed, Guokan Shang et al.

The widespread use of large language models (LLMs) raises an important question: how do texts evolve when they are repeatedly processed by LLMs? In this paper, we define this iterative inference process as Markovian generation chains, where each step takes a specific prompt template and the previous output as input, without including any prior memory. In iterative rephrasing and round-trip translation experiments, the output either converges to a small recurrent set or continues to produce novel sentences over a finite horizon. Through sentence-level Markov chain modeling and analysis of simulated data, we show that iterative process can either increase or reduce sentence diversity depending on factors such as the temperature parameter and the initial input sentence. These results offer valuable insights into the dynamics of iterative LLM inference and their implications for multi-agent LLM systems.

Mohamed Ouf, Amr Mohamed, Mariam Guizani

Open Source Software (OSS) sustainability relies on newcomers transitioning to core contributors, but this pipeline is broken, with most newcomers becoming inactive after initial contributions. Open Source Software for Social Good (OSS4SG) projects, which prioritize societal impact as their primary mission, may be associated with different newcomer-to-core transition outcomes than conventional OSS projects. We compared 375 projects (190 OSS4SG, 185 OSS), analyzing 92,721 contributors and 3.5 million commits. OSS4SG projects retain contributors at 2.2X higher rates and contributors have 19.6% higher probability of achieving core status. Early broad project exploration predicts core achievement (22.2% importance); conventional OSS concentrates on one dominant pathway (61.62% of transitions) while OSS4SG provides multiple pathways. Contrary to intuition, contributors who invest time learning the project before intensifying their contributions (Late Spike pattern) achieve core status 2.4-2.9X faster (21 weeks) than those who contribute intensively from day one (Early Spike pattern, 51-60 weeks). OSS4SG supports two effective temporal patterns while only Late Spike achieves fastest time-to-core in conventional OSS. Our findings suggest that finding a project aligned with personal values and taking time to understand the codebase before major contributions are key strategies for achieving core status. Our findings show that project mission is associated with measurably different environments for newcomer-to-core transitions and provide evidence-based guidance for newcomers and maintainers.

Layth Hamad, Muhammad Asif Khan, Amr Mohamed

Autonomous robots use simultaneous localization and mapping (SLAM) for efficient and safe navigation in various environments. LiDAR sensors are integral in these systems for object identification and localization. However, LiDAR systems though effective in detecting solid objects (e.g., trash bin, bottle, etc.), encounter limitations in identifying semitransparent or non-tangible objects (e.g., fire, smoke, steam, etc.) due to poor reflecting characteristics. Additionally, LiDAR also fails to detect features such as navigation signs and often struggles to detect certain hazardous materials that lack a distinct surface for effective laser reflection. In this paper, we propose a highly accurate stereo-vision approach to complement LiDAR in autonomous robots. The system employs advanced stereo vision-based object detection to detect both tangible and non-tangible objects and then uses simple machine learning to precisely estimate the depth and size of the object. The depth and size information is then integrated into the SLAM process to enhance the robot's navigation capabilities in complex environments. Our evaluation, conducted on an autonomous robot equipped with LiDAR and stereo-vision systems demonstrates high accuracy in the estimation of an object's depth and size. A video illustration of the proposed scheme is available at: \url{https://www.youtube.com/watch?v=nusI6tA9eSk}.

Amr Mohamed, Yang Zhang, Michalis Vazirgiannis et al.

Code-switching (CSW) is the act of alternating between two or more languages within a single discourse. This phenomenon is widespread in multilingual communities, and increasingly prevalent in online content, where users naturally mix languages in everyday communication. As a result, Large Language Models (LLMs), now central to content processing and generation, are frequently exposed to code-switched inputs. Given their widespread use, it is crucial to understand how LLMs process and reason about such mixed-language text. This paper presents a systematic evaluation of LLM comprehension under code-switching by generating CSW variants of established reasoning and comprehension benchmarks. While degradation is evident when foreign tokens disrupt English text$\unicode{x2013}$even under linguistic constraints$\unicode{x2013}$embedding English into other languages often improves comprehension. Though prompting yields mixed results, fine-tuning offers a more stable path to degradation mitigation.

Yang Zhang, Amr Mohamed, Hadi Abdine et al.

Curriculum learning has shown promise in improving training efficiency and generalization in various machine learning domains, yet its potential in pretraining language models remains underexplored, prompting our work as the first systematic investigation in this area. We experimented with different settings, including vanilla curriculum learning, pacing-based sampling, and interleaved curricula-guided by six difficulty metrics spanning linguistic and information-theoretic perspectives. We train models under these settings and evaluate their performance on eight diverse benchmarks. Our experiments reveal that curriculum learning consistently improves convergence in early and mid-training phases, and can yield lasting gains when used as a warmup strategy with up to $3.5\%$ improvement. Notably, we identify compression ratio, lexical diversity, and readability as effective difficulty signals across settings. Our findings highlight the importance of data ordering in large-scale pretraining and provide actionable insights for scalable, data-efficient model development under realistic training scenarios.

Amr Mohamed, Mahmoud Rabea, Aya Sameh et al.

The RSNA-MICCAI brain tumor radiogenomic classification challenge aimed to predict MGMT biomarker status in glioblastoma through binary classification on Multi parameter mpMRI scans: T1w, T1wCE, T2w and FLAIR. The dataset is splitted into three main cohorts: training set, validation set which were used during training, and the testing were only used during final evaluation. Images were either in a DICOM format or in Png format. different architectures were used to investigate the problem including the 3D version of Vision Transformer (ViT3D), ResNet50, Xception and EfficientNet-B3. AUC was used as the main evaluation metric and the results showed an advantage for both the ViT3D and the Xception models achieving 0.6015 and 0.61745 respectively on the testing set. compared to other results, our results proved to be valid given the complexity of the task. further improvements can be made through exploring different strategies, different architectures and more diverse datasets.

Amr Mohamed, Mingmeng Geng, Michalis Vazirgiannis et al.

As large language models are increasingly responsible for online content, concerns arise about the impact of repeatedly processing their own outputs. Inspired by the "broken telephone" effect in chained human communication, this study investigates whether LLMs similarly distort information through iterative generation. Through translation-based experiments, we find that distortion accumulates over time, influenced by language choice and chain complexity. While degradation is inevitable, it can be mitigated through strategic prompting techniques. These findings contribute to discussions on the long-term effects of AI-mediated information propagation, raising important questions about the reliability of LLM-generated content in iterative workflows.

Layth Hamad, Muhammad Asif Khan, Hamid Menouar et al.

This paper presents Haris, an advanced autonomous mobile robot system for tracking the location of vehicles in crowded car parks using license plate recognition. The system employs simultaneous localization and mapping (SLAM) for autonomous navigation and precise mapping of the parking area, eliminating the need for GPS dependency. In addition, the system utilizes a sophisticated framework using computer vision techniques for object detection and automatic license plate recognition (ALPR) for reading and associating license plate numbers with location data. This information is subsequently synchronized with a back-end service and made accessible to users via a user-friendly mobile app, offering effortless vehicle location and alleviating congestion within the parking facility. The proposed system has the potential to improve the management of short-term large outdoor parking areas in crowded places such as sports stadiums. The demo of the robot can be found on https://youtu.be/ZkTCM35fxa0?si=QjggJuN7M1o3oifx.

Alaa Awad Abdellatif, Ali Elmancy, Amr Mohamed et al.

This paper introduces a comprehensive framework for Post-Disaster Search and Rescue (PDSR), aiming to optimize search and rescue operations leveraging Unmanned Aerial Vehicles (UAVs). The primary goal is to improve the precision and availability of sensing capabilities, particularly in various catastrophic scenarios. Central to this concept is the rapid deployment of UAV swarms equipped with diverse sensing, communication, and intelligence capabilities, functioning as an integrated system that incorporates multiple technologies and approaches for efficient detection of individuals buried beneath rubble or debris following a disaster. Within this framework, we propose architectural solution and address associated challenges to ensure optimal performance in real-world disaster scenarios. The proposed framework aims to achieve complete coverage of damaged areas significantly faster than traditional methods using a multi-tier swarm architecture. Furthermore, integrating multi-modal sensing data with machine learning for data fusion could enhance detection accuracy, ensuring precise identification of survivors.

Menna Helmy, Alaa Awad Abdellatif, Naram Mhaisen et al.

The forthcoming 6G networks will embrace a new realm of AI-driven services that requires innovative network slicing strategies, namely slicing for AI, which involves the creation of customized network slices to meet Quality of service (QoS) requirements of diverse AI services. This poses challenges due to time-varying dynamics of users' behavior and mobile networks. Thus, this paper proposes an online learning framework to optimize the allocation of computational and communication resources to AI services, while considering their unique key performance indicators (KPIs), such as accuracy, latency, and cost. We define a problem of optimizing the total accuracy while balancing conflicting KPIs, prove its NP-hardness, and propose an online learning framework for solving it in dynamic environments. We present a basic online solution and two variations employing a pre-learning elimination method for reducing the decision space to expedite the learning. Furthermore, we propose a biased decision space subset selection by incorporating prior knowledge to enhance the learning speed without compromising performance and present two alternatives of handling the selected subset. Our results depict the efficiency of the proposed solutions in converging to the optimal decisions, while reducing decision space and improving time complexity.

Amr Mohamed, Maram Assi, Mariam Guizani

Large language model assistants (LLM-assistants) present new opportunities to transform software development. Developers are increasingly adopting these tools across tasks, including coding, testing, debugging, documentation, and design. Yet, despite growing interest, there is no synthesis of how LLM-assistants affect software developer productivity. In this paper, we present a systematic literature review of 37 peer-reviewed studies published between January 2014 and December 2024 that examine this impact. Our analysis reveals that LLM-assistants offer both considerable benefits and critical risks. Commonly reported gains include minimized code search, accelerated development, and the automation of trivial and repetitive tasks. However, studies also highlight concerns around cognitive offloading, reduced team collaboration, and inconsistent effects on code quality. While the majority of studies (92%) adopt a multi-dimensional perspective by examining at least two SPACE dimensions, reflecting increased awareness of the complexity of developer productivity, only 14% extend beyond three dimensions, indicating substantial room for more integrated evaluations. Satisfaction, Performance, and Efficiency are the most frequently investigated dimensions, whereas Communication and Activity remain underexplored. Most studies are exploratory (64%) and methodologically diverse, but lack longitudinal and team-based evaluations. This review surfaces key research gaps and provides recommendations for future research and practice. All artifacts associated with this study are publicly available at https://zenodo.org/records/15788502.

Abdelrahman Soliman, Ahmed Refaey, Aiman Erbad et al.

Intent-based networks (IBNs) are gaining prominence as an innovative technology that automates network operations through high-level request statements, defining what the network should achieve. In this work, we introduce IntAgent, an intelligent intent LLM agent that integrates NWDAF analytics and tools to fulfill the network operator's intents. Unlike previous approaches, we develop an intent tools engine directly within the NWDAF analytics engine, allowing our agent to utilize live network analytics to inform its reasoning and tool selection. We offer an enriched, 3GPP-compliant data source that enhances the dynamic, context-aware fulfillment of network operator goals, along with an MCP tools server for scheduling, monitoring, and analytics tools. We demonstrate the efficacy of our framework through two practical use cases: ML-based traffic prediction and scheduled policy enforcement, which validate IntAgent's ability to autonomously fulfill complex network intents.

Zaineh Abughazzah, Emna Baccour, Loay Ismail et al.

The integration of Unmanned Aerial Vehicles (UAVs) into Open Radio Access Networks (O-RAN) enhances communication in disaster management and Search and Rescue (SAR) operations by ensuring connectivity when infrastructure fails. However, SAR scenarios demand stringent security and low-latency communication, as delays or breaches can compromise mission success. While UAVs serve as mobile relays, they introduce challenges in energy consumption and resource management, necessitating intelligent allocation strategies. Existing UAV-assisted O-RAN approaches often overlook the joint optimization of security, latency, and energy efficiency in dynamic environments. This paper proposes a novel Reinforcement Learning (RL)-based framework for dynamic resource allocation in UAV relays, explicitly addressing these trade-offs. Our approach formulates an optimization problem that integrates security-aware resource allocation, latency minimization, and energy efficiency, which is solved using RL. Unlike heuristic or static methods, our framework adapts in real-time to network dynamics, ensuring robust communication. Simulations demonstrate superior performance compared to heuristic baselines, achieving enhanced security and energy efficiency while maintaining ultra-low latency in SAR scenarios.

Guokan Shang, Hadi Abdine, Ahmad Chamma et al.

We introduce Nile-Chat-4B, 3x4B-A6B, and 12B, a collection of LLMs for Egyptian dialect, uniquely designed to understand and generate texts written in both Arabic and Latin scripts. Specifically, with Nile-Chat-3x4B-A6B, we introduce a novel language adaptation approach by leveraging the Branch-Train-MiX strategy to merge script-specialized experts, into a single MoE model. Our Nile-Chat models significantly outperform leading multilingual and Arabic LLMs, such as LLaMa, Jais, and ALLaM, on our newly introduced Egyptian evaluation benchmarks, which span both understanding and generative tasks. Notably, our 12B model yields a 14.4% performance gain over Qwen2.5-14B-Instruct on Latin-script benchmarks. All our resources are publicly available. We believe this work presents a comprehensive methodology for adapting LLMs to dual-script languages, addressing an often overlooked aspect in modern LLM development.

Abdul Hannaan, Zubair Shah, Aiman Erbad et al.

This paper introduces a novel federated learning framework termed LoRa-FL designed for training low-rank one-shot image detection models deployed on edge devices. By incorporating low-rank adaptation techniques into one-shot detection architectures, our method significantly reduces both computational and communication overhead while maintaining scalable accuracy. The proposed framework leverages federated learning to collaboratively train lightweight image recognition models, enabling rapid adaptation and efficient deployment across heterogeneous, resource-constrained devices. Experimental evaluations on the MNIST and CIFAR10 benchmark datasets, both in an independent-and-identically-distributed (IID) and non-IID setting, demonstrate that our approach achieves competitive detection performance while significantly reducing communication bandwidth and compute complexity. This makes it a promising solution for adaptively reducing the communication and compute power overheads, while not sacrificing model accuracy.

Mhd Saria Allahham, Sameh Sorour, Amr Mohamed et al.

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.

Mhd Saria Allahham, Sameh Sorour, Amr Mohamed et al.

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.

Alaa Awad Abdellatif, Naram Mhaisen, Zina Chkirbene et al.

The rapid increase in the percentage of chronic disease patients along with the recent pandemic pose immediate threats on healthcare expenditure and elevate causes of death. This calls for transforming healthcare systems away from one-on-one patient treatment into intelligent health systems, to improve services, access and scalability, while reducing costs. Reinforcement Learning (RL) has witnessed an intrinsic breakthrough in solving a variety of complex problems for diverse applications and services. Thus, we conduct in this paper a comprehensive survey of the recent models and techniques of RL that have been developed/used for supporting Intelligent-healthcare (I-health) systems. This paper can guide the readers to deeply understand the state-of-the-art regarding the use of RL in the context of I-health. Specifically, we first present an overview for the I-health systems challenges, architecture, and how RL can benefit these systems. We then review the background and mathematical modeling of different RL, Deep RL (DRL), and multi-agent RL models. After that, we provide a deep literature review for the applications of RL in I-health systems. In particular, three main areas have been tackled, i.e., edge intelligence, smart core network, and dynamic treatment regimes. Finally, we highlight emerging challenges and outline future research directions in driving the future success of RL in I-health systems, which opens the door for exploring some interesting and unsolved problems.

Dawei Du, Longyin Wen, Pengfei Zhu et al.

Crowd counting on the drone platform is an interesting topic in computer vision, which brings new challenges such as small object inference, background clutter and wide viewpoint. However, there are few algorithms focusing on crowd counting on the drone-captured data due to the lack of comprehensive datasets. To this end, we collect a large-scale dataset and organize the Vision Meets Drone Crowd Counting Challenge (VisDrone-CC2020) in conjunction with the 16th European Conference on Computer Vision (ECCV 2020) to promote the developments in the related fields. The collected dataset is formed by $3,360$ images, including $2,460$ images for training, and $900$ images for testing. Specifically, we manually annotate persons with points in each video frame. There are $14$ algorithms from $15$ institutes submitted to the VisDrone-CC2020 Challenge. We provide a detailed analysis of the evaluation results and conclude the challenge. More information can be found at the website: \url{http://www.aiskyeye.com/}.

Alaa Awad Abdellatif, Naram Mhaisen, Amr Mohamed et al.

Federated learning (FL) is a distributed learning methodology that allows multiple nodes to cooperatively train a deep learning model, without the need to share their local data. It is a promising solution for telemonitoring systems that demand intensive data collection, for detection, classification, and prediction of future events, from different locations while maintaining a strict privacy constraint. Due to privacy concerns and critical communication bottlenecks, it can become impractical to send the FL updated models to a centralized server. Thus, this paper studies the potential of hierarchical FL in IoT heterogeneous systems and propose an optimized solution for user assignment and resource allocation on multiple edge nodes. In particular, this work focuses on a generic class of machine learning models that are trained using gradient-descent-based schemes while considering the practical constraints of non-uniformly distributed data across different users. We evaluate the proposed system using two real-world datasets, and we show that it outperforms state-of-the-art FL solutions. In particular, our numerical results highlight the effectiveness of our approach and its ability to provide 4-6% increase in the classification accuracy, with respect to hierarchical FL schemes that consider distance-based user assignment. Furthermore, the proposed approach could significantly accelerate FL training and reduce communication overhead by providing 75-85% reduction in the communication rounds between edge nodes and the centralized server, for the same model accuracy.

Emna Baccour, Fatima Haouari, Aiman Erbad et al.

Crowdsourced live video streaming (livecast) services such as Facebook Live, YouNow, Douyu and Twitch are gaining more momentum recently. Allocating the limited resources in a cost-effective manner while maximizing the Quality of Service (QoS) through real-time delivery and the provision of the appropriate representations for all viewers is a challenging problem. In our paper, we introduce a machine-learning based predictive resource allocation framework for geo-distributed cloud sites, considering the delay and quality constraints to guarantee the maximum QoS for viewers and the minimum cost for content providers. First, we present an offline optimization that decides the required transcoding resources in distributed regions near the viewers with a trade-off between the QoS and the overall cost. Second, we use machine learning to build forecasting models that proactively predict the approximate transcoding resources to be reserved at each cloud site ahead of time. Finally, we develop a Greedy Nearest and Cheapest algorithm (GNCA) to perform the resource allocation of real-time broadcasted videos on the rented resources. Extensive simulations have shown that GNCA outperforms the state-of-the art resource allocation approaches for crowdsourced live streaming by achieving more than 20% gain in terms of system cost while serving the viewers with relatively lower latency.

Mohammed Jouhari, Abdulla Al-Ali, Emna Baccour et al.

Unmanned Aerial Vehicles (UAVs) have attracted great interest in the last few years owing to their ability to cover large areas and access difficult and hazardous target zones, which is not the case of traditional systems relying on direct observations obtained from fixed cameras and sensors. Furthermore, thanks to the advancements in computer vision and machine learning, UAVs are being adopted for a broad range of solutions and applications. However, Deep Neural Networks (DNNs) are progressing toward deeper and complex models that prevent them from being executed on-board. In this paper, we propose a DNN distribution methodology within UAVs to enable data classification in resource-constrained devices and avoid extra delays introduced by the server-based solutions due to data communication over air-to-ground links. The proposed method is formulated as an optimization problem that aims to minimize the latency between data collection and decision-making while considering the mobility model and the resource constraints of the UAVs as part of the air-to-air communication. We also introduce the mobility prediction to adapt our system to the dynamics of UAVs and the network variation. The simulation conducted to evaluate the performance and benchmark the proposed methods, namely Optimal UAV-based Layer Distribution (OULD) and OULD with Mobility Prediction (OULD-MP), were run in an HPC cluster. The obtained results show that our optimization solution outperforms the existing and heuristic-based approaches.

Emna Baccour, Naram Mhaisen, Alaa Awad Abdellatif et al.

Artificial intelligence (AI) has witnessed a substantial breakthrough in a variety of Internet of Things (IoT) applications and services, spanning from recommendation systems to robotics control and military surveillance. This is driven by the easier access to sensory data and the enormous scale of pervasive/ubiquitous devices that generate zettabytes (ZB) of real-time data streams. Designing accurate models using such data streams, to predict future insights and revolutionize the decision-taking process, inaugurates pervasive systems as a worthy paradigm for a better quality-of-life. The confluence of pervasive computing and artificial intelligence, Pervasive AI, expanded the role of ubiquitous IoT systems from mainly data collection to executing distributed computations with a promising alternative to centralized learning, presenting various challenges. In this context, a wise cooperation and resource scheduling should be envisaged among IoT devices (e.g., smartphones, smart vehicles) and infrastructure (e.g. edge nodes, and base stations) to avoid communication and computation overheads and ensure maximum performance. In this paper, we conduct a comprehensive survey of the recent techniques developed to overcome these resource challenges in pervasive AI systems. Specifically, we first present an overview of the pervasive computing, its architecture, and its intersection with artificial intelligence. We then review the background, applications and performance metrics of AI, particularly Deep Learning (DL) and online learning, running in a ubiquitous system. Next, we provide a deep literature review of communication-efficient techniques, from both algorithmic and system perspectives, of distributed inference, training and online learning tasks across the combination of IoT devices, edge devices and cloud servers. Finally, we discuss our future vision and research challenges.

Ali Hamdi, Khaled Shaban, Abdelkarim Erradi et al.

Spatiotemporal data mining (STDM) discovers useful patterns from the dynamic interplay between space and time. Several available surveys capture STDM advances and report a wealth of important progress in this field. However, STDM challenges and problems are not thoroughly discussed and presented in articles of their own. We attempt to fill this gap by providing a comprehensive literature survey on state-of-the-art advances in STDM. We describe the challenging issues and their causes and open gaps of multiple STDM directions and aspects. Specifically, we investigate the challenging issues in regards to spatiotemporal relationships, interdisciplinarity, discretisation, and data characteristics. Moreover, we discuss the limitations in the literature and open research problems related to spatiotemporal data representations, modelling and visualisation, and comprehensiveness of approaches. We explain issues related to STDM tasks of classification, clustering, hotspot detection, association and pattern mining, outlier detection, visualisation, visual analytics, and computer vision tasks. We also highlight STDM issues related to multiple applications including crime and public safety, traffic and transportation, earth and environment monitoring, epidemiology, social media, and Internet of Things.

Naram Mhaisen, Alaa Awad, Amr Mohamed et al.

Distributed learning algorithms aim to leverage distributed and diverse data stored at users' devices to learn a global phenomena by performing training amongst participating devices and periodically aggregating their local models' parameters into a global model. Federated learning is a promising paradigm that allows for extending local training among the participant devices before aggregating the parameters, offering better communication efficiency. However, in the cases where the participants' data are strongly skewed (i.e., non-IID), the local models can overfit local data, leading to low performing global model. In this paper, we first show that a major cause of the performance drop is the weighted distance between the distribution over classes on users' devices and the global distribution. Then, to face this challenge, we leverage the edge computing paradigm to design a hierarchical learning system that performs Federated Gradient Descent on the user-edge layer and Federated Averaging on the edge-cloud layer. In this hierarchical architecture, we formalize and optimize this user-edge assignment problem such that edge-level data distributions turn to be similar (i.e., close to IID), which enhances the Federated Averaging performance. Our experiments on multiple real-world datasets show that the proposed optimized assignment is tractable and leads to faster convergence of models towards a better accuracy value.

Alaa Awad Abdellatif, Abeer Z. Al-Marridi, Amr Mohamed et al.

The future of healthcare systems is being shaped by incorporating emerged technological innovations to drive new models for patient care. By acquiring, integrating, analyzing, and exchanging medical data at different system levels, new practices can be introduced, offering a radical improvement to healthcare services. This paper presents a novel smart and secure Healthcare system (ssHealth), which, leveraging advances in edge computing and blockchain technologies, permits epidemics discovering, remote monitoring, and fast emergency response. The proposed system also allows for secure medical data exchange among local healthcare entities, thus realizing the integration of multiple national and international entities and enabling the correlation of critical medical events for, e.g., emerging epidemics management and control. In particular, we develop a blockchain-based architecture and enable a flexible configuration thereof, which optimize medical data sharing between different health entities and fulfil the diverse levels of Quality of Service (QoS) that ssHealth may require. Finally, we highlight the benefits of the proposed ssHealth system and possible directions for future research.

Emna Baccour, Aiman Erbad, Kashif Bilal et al.

With the advancement in personal smart devices and pervasive network connectivity, users are no longer passive content consumers, but also contributors in producing new contents. This expansion in live services requires a detailed analysis of broadcasters' and viewers' behavior to maximize users' Quality of Experience (QoE). In this paper, we present a dataset gathered from one of the popular live streaming platforms: Facebook. In this dataset, we stored more than 1,500,000 live stream records collected in June and July 2018. These data include public live videos from all over the world. However, Facebook live API does not offer the possibility to collect online videos with their fine grained data. The API allows to get the general data of a stream, only if we know its ID (identifier). Therefore, using the live map website provided by Facebook and showing the locations of online streams and locations of viewers, we extracted video IDs and different coordinates along with general metadata. Then, having these IDs and using the API, we can collect the fine grained metadata of public videos that might be useful for the research community. We also present several preliminary analyses to describe and identify the patterns of the streams and viewers. Such fine grained details will enable the multimedia community to recreate real-world scenarios particularly for resource allocation, caching, computation, and transcoding in edge networks. Existing datasets do not provide the locations of the viewers, which limits the efforts made to allocate the multimedia resources as close as possible to viewers and to offer better QoE.

Fatima Haouari, Emna Baccour, Aiman Erbad et al.

Driven by the tremendous technological advancement of personal devices and the prevalence of wireless mobile network accesses, the world has witnessed an explosion in crowdsourced live streaming. Ensuring a better viewers quality of experience (QoE) is the key to maximize the audiences number and increase streaming providers' profits. This can be achieved by advocating a geo-distributed cloud infrastructure to allocate the multimedia resources as close as possible to viewers, in order to minimize the access delay and video stalls. Moreover, allocating the exact needed resources beforehand avoids over-provisioning, which may lead to significant costs by the service providers. In the contrary, under-provisioning might cause significant delays to the viewers. In this paper, we introduce a prediction driven resource allocation framework, to maximize the QoE of viewers and minimize the resource allocation cost. First, by exploiting the viewers locations available in our unique dataset, we implement a machine learning model to predict the viewers number near each geo-distributed cloud site. Second, based on the predicted results that showed to be close to the actual values, we formulate an optimization problem to proactively allocate resources at the viewers proximity. Additionally, we will present a trade-off between the video access delay and the cost of resource allocation.

Emna Baccour, Aiman Erbad, Amr Mohamed et al.

Recently, the growing demand for rich multimedia content such as Video on Demand (VoD) has made the data transmission from content delivery networks (CDN) to end-users quite challenging. Edge networks have been proposed as an extension to CDN networks to alleviate this excessive data transfer through caching and to delegate the computation tasks to edge servers. To maximize the caching efficiency in the edge networks, different Mobile Edge Computing (MEC) servers assist each others to effectively select which content to store and the appropriate computation tasks to process. In this paper, we adopt a collaborative caching and transcoding model for VoD in MEC networks. However, unlike other models in the literature, different chunks of the same video are not fetched and cached in the same MEC server. Instead, neighboring servers will collaborate to store and transcode different video chunks and consequently optimize the limited resources usage. Since we are dealing with chunks caching and processing, we propose to maximize the edge efficiency by studying the viewers watching pattern and designing a probabilistic model where chunks popularities are evaluated. Based on this model, popularity-aware policies, namely Proactive caching policy (PcP) and Cache replacement Policy (CrP), are introduced to cache only highest probably requested chunks. In addition to PcP and CrP, an online algorithm (PCCP) is proposed to schedule the collaborative caching and processing. The evaluation results prove that our model and policies give better performance than approaches using conventional replacement policies. This improvement reaches up to 50% in some cases.

Reza Shakeri, Mohammed Ali Al-Garadi, Ahmed Badawy et al.

Unmanned Aerial Vehicles (UAVs) have recently rapidly grown to facilitate a wide range of innovative applications that can fundamentally change the way cyber-physical systems (CPSs) are designed. CPSs are a modern generation of systems with synergic cooperation between computational and physical potentials that can interact with humans through several new mechanisms. The main advantages of using UAVs in CPS application is their exceptional features, including their mobility, dynamism, effortless deployment, adaptive altitude, agility, adjustability, and effective appraisal of real-world functions anytime and anywhere. Furthermore, from the technology perspective, UAVs are predicted to be a vital element of the development of advanced CPSs. Therefore, in this survey, we aim to pinpoint the most fundamental and important design challenges of multi-UAV systems for CPS applications. We highlight key and versatile aspects that span the coverage and tracking of targets and infrastructure objects, energy-efficient navigation, and image analysis using machine learning for fine-grained CPS applications. Key prototypes and testbeds are also investigated to show how these practical technologies can facilitate CPS applications. We present and propose state-of-the-art algorithms to address design challenges with both quantitative and qualitative methods and map these challenges with important CPS applications to draw insightful conclusions on the challenges of each application. Finally, we summarize potential new directions and ideas that could shape future research in these areas.

Mohammed Ali Al-Garadi, Amr Mohamed, Abdulla Al-Ali et al.

The Internet of Things (IoT) integrates billions of smart devices that can communicate with one another with minimal human intervention. It is one of the fastest developing fields in the history of computing, with an estimated 50 billion devices by the end of 2020. On the one hand, IoT play a crucial role in enhancing several real-life smart applications that can improve life quality. On the other hand, the crosscutting nature of IoT systems and the multidisciplinary components involved in the deployment of such systems introduced new security challenges. Implementing security measures, such as encryption, authentication, access control, network security and application security, for IoT devices and their inherent vulnerabilities is ineffective. Therefore, existing security methods should be enhanced to secure the IoT system effectively. Machine learning and deep learning (ML/DL) have advanced considerably over the last few years, and machine intelligence has transitioned from laboratory curiosity to practical machinery in several important applications. Consequently, ML/DL methods are important in transforming the security of IoT systems from merely facilitating secure communication between devices to security-based intelligence systems. The goal of this work is to provide a comprehensive survey of ML /DL methods that can be used to develop enhanced security methods for IoT systems. IoT security threats that are related to inherent or newly introduced threats are presented, and various potential IoT system attack surfaces and the possible threats related to each surface are discussed. We then thoroughly review ML/DL methods for IoT security and present the opportunities, advantages and shortcomings of each method. We discuss the opportunities and challenges involved in applying ML/DL to IoT security. These opportunities and challenges can serve as potential future research directions.

Chenglong Fu, Xiaojiang Du, Longfei Wu et al.

Implantable medical devices (IMDs), such as pacemakers, implanted cardiac defibrillators, and neurostimulators are medical devices implanted into patients' bodies for monitoring physiological signals and performing medical treatments. Many IMDs have built-in wireless communication modules to facilitate data collecting and device reprogramming by external programmers. The wireless communication brings significant conveniences for advanced applications such as real-time and remote monitoring but also introduces the risk of unauthorized wireless access. The absence of effective access control mechanisms exposes patients' life to cyber attacks. In this paper, we present a lightweight and universally applicable access control system for IMDs. By leveraging Physically Obfuscated Keys (POKs) as the hardware root of trust, provable security is achieved based on standard cryptographic primitives while attaining high energy efficiency. In addition, barrier-free IMD access under emergent situations is realized by utilizing the patient's biometrical information. We evaluate our proposed scheme through extensive security analysis and a prototype implementation, which demonstrates our work's superiority on security and energy efficiency.

Ahmed Ben Said, Amr Mohamed, Tarek Elfouly et al.

In this paper, we present a joint compression and classification approach of EEG and EMG signals using a deep learning approach. Specifically, we build our system based on the deep autoencoder architecture which is designed not only to extract discriminant features in the multimodal data representation but also to reconstruct the data from the latent representation using encoder-decoder layers. Since autoencoder can be seen as a compression approach, we extend it to handle multimodal data at the encoder layer, reconstructed and retrieved at the decoder layer. We show through experimental results, that exploiting both multimodal data intercorellation and intracorellation 1) Significantly reduces signal distortion particularly for high compression levels 2) Achieves better accuracy in classifying EEG and EMG signals recorded and labeled according to the sentiments of the volunteer.

Ahmed Saeed, Ahmed Abdelkader, Mouhyemen Khan et al.

Low-cost mini-drones with advanced sensing and maneuverability enable a new class of intelligent sensing systems. To achieve the full potential of such drones, it is necessary to develop new enhanced formulations of both common and emerging sensing scenarios. Namely, several fundamental challenges in visual sensing remain unsolved including: 1) Fitting sizable targets in camera frames; 2) Effective viewpoints matching target poses; 3) Occlusion by elements in the environment, including other targets. In this paper, we introduce Argus: an autonomous system that utilizes drones to incrementally collect target information through a two-tier architecture. To tackle the stated challenges, Argus employs a novel geometric model that captures both target shapes and coverage constraints. Recognizing drones as the scarcest resource, Argus aims to minimize the number of drones required to cover a set of targets. We prove this problem is NP-hard, and even hard to approximate, before deriving a best-possible approximation algorithm along with a competitive sampling heuristic which runs up to 100x faster according to large-scale simulations. To test Argus in action, we demonstrate and analyze its performance on a prototype implementation. Finally, we present a number of extensions to accommodate more application requirements and highlight some open problems.

Abduljalil Mohamed, Khaled Bashir Shaban, Amr Mohamed

Wireless electroencephalogram (EEG) sensors have been successfully applied in many medical and computer brain interface classifications. A common characteristic of wireless EEG sensors is that they are low powered devices, and hence an efficient usage of sensor energy resources is critical for any practical application. One way of minimizing energy consumption by the EEG sensors is by reducing the number of EEG channels participating in the classification process. For the purpose of classifying EEG signals, we propose a directed acyclic graph (DAG)-based channel selection algorithm. To achieve this objective, the EEG sensor channels are first realized in a complete undirected graph, where each channel is represented by a node. An edge between any two nodes indicates the collaboration between these nodes in identifying the system state; and the significance of this collaboration is quantified by a weight assigned to the edge. The complete graph is then reduced into a directed acyclic graph that encodes the knowledge of the non-increasing order of the channel ranking for each cognitive task. The channel selection algorithm utilizes this directed graph to find a maximum path such that the total weight of this path satisfies a predefined threshold. It has been demonstrated experimentally that channel utilization has been reduced by 50% in the worst case scenario for a three-state system and an EEG sensor with 14 channels; and the best classification accuracy obtained is 81%.

Ahmed El Shafie, Tamer Khattab, Hussien Saad et al.

In this paper, we consider a cognitive setting under the context of cooperative communications, where the cognitive radio (CR) user is assumed to be a self-organized relay for the network. The CR user and the PU are assumed to be energy harvesters. The CR user cooperatively relays some of the undelivered packets of the primary user (PU). Specifically, the CR user stores a fraction of the undelivered primary packets in a relaying queue (buffer). It manages the flow of the undelivered primary packets to its relaying queue using the appropriate actions over time slots. Moreover, it has the decision of choosing the used queue for channel accessing at idle time slots (slots where the PU's queue is empty). It is assumed that one data packet transmission dissipates one energy packet. The optimal policy changes according to the primary and CR users arrival rates to the data and energy queues as well as the channels connectivity. The CR user saves energy for the PU by taking the responsibility of relaying the undelivered primary packets. It optimally organizes its own energy packets to maximize its payoff as time progresses.

Hussein Saad, Amr Mohamed, Tamer ElBatt

In this paper, we address the problem of distributed interference management of cognitive femtocells that share the same frequency range with macrocells (primary user) using distributed multi-agent Q-learning. We formulate and solve three problems representing three different Q-learning algorithms: namely, centralized, distributed and partially distributed power control using Q-learning (CPC-Q, DPC-Q and PDPC-Q). CPCQ, although not of practical interest, characterizes the global optimum. Each of DPC-Q and PDPC-Q works in two different learning paradigms: Independent (IL) and Cooperative (CL). The former is considered the simplest form for applying Qlearning in multi-agent scenarios, where all the femtocells learn independently. The latter is the proposed scheme in which femtocells share partial information during the learning process in order to strike a balance between practical relevance and performance. In terms of performance, the simulation results showed that the CL paradigm outperforms the IL paradigm and achieves an aggregate femtocells capacity that is very close to the optimal one. For the practical relevance issue, we evaluate the robustness and scalability of DPC-Q, in real time, by deploying new femtocells in the system during the learning process, where we showed that DPC-Q in the CL paradigm is scalable to large number of femtocells and more robust to the network dynamics compared to the IL paradigm

Hussein Saad, Amr Mohamed, Tamer ElBatt

In this paper, we propose a distributed reinforcement learning (RL) technique called distributed power control using Q-learning (DPC-Q) to manage the interference caused by the femtocells on macro-users in the downlink. The DPC-Q leverages Q-Learning to identify the sub-optimal pattern of power allocation, which strives to maximize femtocell capacity, while guaranteeing macrocell capacity level in an underlay cognitive setting. We propose two different approaches for the DPC-Q algorithm: namely, independent, and cooperative. In the former, femtocells learn independently from each other while in the latter, femtocells share some information during learning in order to enhance their performance. Simulation results show that the independent approach is capable of mitigating the interference generated by the femtocells on macro-users. Moreover, the results show that cooperation enhances the performance of the femtocells in terms of speed of convergence, fairness and aggregate femtocell capacity.