The pervasive issue of degree fraud poses a signif-icant challenge to the credibility of academic credentials, neces-sitating innovative solutions to ensure the integrity of diploma verification processes. This paper proposes an integrated solution leveraging Blockchain, InterPlanetary File System (IPFS), and Advanced Encryption Standard (AES) encryption technologies to establish a secure and transparent diploma management system. IPFS offers a decentralized approach to file storage and distribution, eliminating reliance on centralized authorities and enhancing data availability and accessibility. Similarly, Blockchain technology provides a tamper-resistant ledger for storing diploma data, ensuring transparency and immutability through distributed consensus mechanisms. Ethereum, with its smart contract functionality, facilitates the automation of trans-actional processes, further enhancing the system's efficiency and reliability, finally, AES ensures the confidentiality of certificate contents through strong encryption, preventing unauthorized access to sensitive information. The loading process for university diplomas involves encrypting PDF files with AES encryption, uploading them to the IPFS cluster, and recording their unique hash values on the Ethereum blockchain. Subsequently, the verification process requires accurate specification of both the file and transaction hash values, ensuring the authenticity of diplomas through blockchain-based verification. The implemen-tation of the proposed system demonstrates its efficacy in securely uploading and verifying diplomas, utilizing technologies such as MetaMask for Ethereum authentication and AES encryption for data confidentiality. By ensuring the integrity and authenticity of diplomas, the system mitigates the risk of fraud and upholds the integrity of educational qualifications.
The literature review shows the need for Key Exchange Protocols and Message Authentication Codes in secure communication protocols such as SSH, IPsec, and TLS. Nevertheless, the hardness of the Learning With Errors (LWE) problem and the security of SHA3 hashes have been proven critical to developing secure and quantum-safe cryptographic schemes, which are critical in this digital era. Indeed, current privacy laws and regulations worldwide require protecting citizens' personally identifiable data in different sectors such as healthcare, finance, and e-government. This paper proposes new asymmetric cryptographic schemes for key exchange and new symmetric ones for authentication. My Key Exchange Protocols rely on mathematical concepts such as the newly defined “new paradigm related to the learning with errors” problem and the symmetric matrix. In addition, my new Message Authentication Codes rely on the aforementioned hard problem and the secure hash function SHA3. This paper provides correctness, security, quantum safety, and efficiency proofs of my cryptographic schemes. It is important to highlight the urgency of developing quantum-safe protocols. Indeed, with the growth of the storage capacity, when democratized, quantum computers can threaten the currently stored data.
Brain tumor is one of the most frequently diagnosed types of cancer, and understanding the intricacies of the brain is highly intricate and vital, leading to extensive research efforts. In the realm of artificial neural networks (ANNs), particularly in pattern recognition tasks, methods such as feed-forward neural networks (FNNs) are utilized due to their ability to learn complex patterns and handle intricate data structures. However, the efficacy of these networks heavily relies on their architecture and the optimization algorithms used to fine-tune their parameters. In this study, novel yet efficient methodologies for training artificial neural networks are introduced. The majority of contemporary training techniques for artificial neural networks are based on gradient descent algorithms. In contrast to conventional approaches, we propose an advancement by integrating the Evolutionary Algorithm for Brain Tumor Prediction into the training process. Additionally, we explore the utilization of other optimization algorithms such as Genetic Algorithm, alongside conventional methods like Adam. Our research involves the development of an artificial neural network model tailored for pattern recognition tasks, training the model using various optimization algorithms, and evaluating its performance through extensive experiments on the Brain MRI dataset.
Recent advancements in Software-Defined Networking (SDN) have facilitated its deployment across diverse network types, including edge networks. Given the broad applicability of SDN and the complexity of large-scale environments, establishing a comprehensive real-world test environment is both challenging and expensive. To circumvent these obstacles, software-based simulations are typically employed to validate solutions prior to real-world deployment. However, these simulations often do not incorporate real-time hardware data, limiting their realism. This paper introduces a novel hybrid SDN simulation testbed that integrates real hardware data within a Mininet-emulated network, addressing this limitation. To demonstrate the efficacy of our hybrid testbed, we present a specific scenario involving the dynamic allocation of edge resources to various client requests through a machine learning approach. This scenario focuses on detecting LiDAR spoofing attacks within automotive systems. Additionally, our hybrid testbed facilitates the generation and replication of new datasets for tailored scenarios, enhancing research capabilities in more intricate contexts.
Indoor Positioning Systems (IPS) play a crucial role in various applications, including asset tracking, navigation, and context-aware services. They rely on technologies such as WiFi, Bluetooth, and Ultra-Wideband (UWB) to determine the location of devices within enclosed spaces. By leveraging Mobile Access Points (MAP) and innovative algorithms, this paper aims to overcome the limitations of traditional IPS methods, particularly in challenging environments with Non- Line of Sight (NLOS) conditions. Comparisons are made with existing approaches, including the fingerprinting technique for signal strength analysis. Results demonstrate the effectiveness of the proposed system, achieving an impressive RMSE of 0.26 m. This work contributes to bridging the gap between telecom and data processing elements, translating simulations into real-world experiences, and providing a comprehensive evaluation framework.
