Institution profile

Khalifa University

Academic institutionnorthamerica · us
Official website
Research library276linked papers
Opportunities0open roles
Selected work

Representative Papers

Influence- and Interest-Based Worker Recruitment in Crowdsourcing Using Online Social Networks

Jun 01, 2023IEEE Transactions on Network and Service Management

To address the cold-start problem in mobile crowdsourcing—characterized by an initial scarcity of workers and low task response rates (~4%)—and to overcome limitations of existing recruitment methods in team composition, procedural clarity, and adaptability, this paper proposes an influencer-driven recruitment framework leveraging online social networks. Our method introduces a grouped, interest-aware influence maximization algorithm and a dynamic backup mechanism that replaces workers who decline tasks in real time, enabling end-to-end executable recruitment. It integrates social attribute modeling, genetic optimization, and dynamic task allocation. Extensive evaluation on real-world datasets demonstrates that our approach improves effective worker coverage by 32.7% on average over baseline methods, significantly enhancing Quality-of-Service (QoS) guarantees.

4 citationsRead paper

Enabling AutoML for Zero-Touch Network Security: Use-Case Driven Analysis

Jun 01, 2024IEEE Transactions on Network and Service Management

In 6G zero-touch networks (ZTNs), AI/ML-based security mechanisms suffer from labor-intensive hyperparameter tuning and vulnerability to adversarial attacks, posing critical safety bottlenecks. Method: This work proposes the first AutoML-empowered fully automated security framework for ZTNs, integrating automated machine learning (AutoML), adversarial machine learning (AML), multi-source anomaly detection, and explainable AI (XAI) to jointly model network traffic and model behavior—enabling autonomous intrusion detection and integrated adversarial defense. Contribution/Results: Experimental evaluation across multiple representative scenarios demonstrates >98.5% intrusion detection accuracy and robust resistance against mainstream adversarial attacks, including FGSM and PGD. The framework delivers the first deployable end-to-end AutoML solution for ZTN security, significantly reducing human intervention while enhancing model robustness and operational autonomy.

3 citationsRead paper

Blockchain-Assisted Demonstration Cloning for Multiagent Deep Reinforcement Learning

Mar 01, 2024IEEE Internet of Things Journal

To address low sample efficiency, the curse of dimensionality, and insufficient exploration in multi-agent deep reinforcement learning (MDRL), this paper proposes a blockchain-empowered Multi-Expert Demonstration Cloning (MEDC) framework. MEDC innovatively integrates a permissioned blockchain with IPFS to enable trustworthy sharing and end-to-end provenance tracking of expert models; it leverages smart contracts to automatically distribute verified expert policies and augments multi-agent DRL with imitation learning—thereby circumventing malicious model injection risks inherent in federated learning and avoiding local optima induced by reward shaping. Experiments across multiple benchmark tasks demonstrate that MEDC significantly improves convergence speed and robustness, exhibits strong fault tolerance against model failures and adversarial policy injection, and consistently outperforms state-of-the-art federated RL, reward shaping, and imitation-learning-augmented approaches.

2 citationsRead paper

Blockchain-based crowdsourced deep reinforcement learning as a service

Jun 01, 2024Information Sciences

To address the high adoption barrier of deep reinforcement learning (DRL) and the lack of trustworthy mechanisms for collaborative training, this paper proposes the first blockchain-enabled decentralized DRL crowdsourced training framework. The framework integrates a permissioned blockchain (Hyperledger Fabric), federated reinforcement learning, differential privacy, and smart contracts to ensure verifiable training, traceable accountability, and fair reward distribution. Evaluated on CartPole and traffic signal control tasks, it achieves a 23% faster policy convergence compared to conventional federated learning and improves model robustness by 37% under adversarial participant attacks. Its core contributions are: (i) introducing the first DRL crowdsourced training paradigm; (ii) resolving trust bottlenecks and the privacy-utility trade-off in multi-party collaboration; and (iii) providing a novel pathway toward scalable, secure, and production-ready DRL deployment.

1 citationsRead paper
Recent publications

Latest Papers