Secure blockchain integrated deep learning framework for federated risk-adaptive and privacy-preserving IoT edge...

Area of Science:

• Computer Science
• Artificial Intelligence
• Cybersecurity

Background:

• The proliferation of Internet of Things (IoT) devices necessitates secure, scalable, and intelligent edge computing frameworks.
• Edge nodes, physically deployed in vulnerable environments, face risks of resource manipulation and cyber threats.
• Existing solutions often fail to balance security with resource constraints, impacting IoT system robustness.

Purpose of the Study:

• To propose a novel framework integrating blockchain and deep learning for secure IoT edge computing.
• To address the limitations of current edge computing security strategies by enhancing trustworthiness and adaptability.
• To develop a hybrid architecture that leverages blockchain's transparency and deep learning's flexibility.

Main Methods:

• Blockchain-Orchestrated Federated Curriculum Learning (BOFCL) for risk-prioritized training using blockchain threat indices.
• Zero-Knowledge Proof Enabled Secure Inference Engine (ZK-SIE) for privacy-preserving and verifiable model inference.
• Blockchain Indexed Adversarial Attack Simulator (BI-AAS) for testing and retraining models against simulated attacks.
• Energy-Aware Lightweight Consensus with Adaptive Synchronization (ELCAS) for efficient global model synchronization.
• Trust Indexed Model Provenance and Deployment Ledger (TIMPDL) for transparent model lineage tracking and deployment.

Main Results:

• The framework combines data integrity, adversarial robustness, and trust-aware deployment for edge intelligence.
• Achieved reductions in training latency, synchronization energy consumption, and privacy leakage.
• Enhanced responsiveness to high-risk edge scenarios through adaptive training sequencing.
• Ensured verifiable privacy-preserving inference and model integrity without data exposure.

Conclusions:

• The proposed Blockchain Integrated Deep Learning Framework offers a foundational advancement for secure decentralized edge intelligence in IoT.
• The hybrid architecture successfully balances security requirements with the resource constraints of edge environments.
• This approach significantly improves the robustness and trustworthiness of IoT edge computing solutions.