Journal of Ambient Intelligence and Humanized Computing最新文献

Gene expression profiling for cancer diagnosis requires the identification of optimal and non-redundant gene subsets from microarray data. We present a multi-objective particle swarm optimization (PSO) approach that balances gene-class relevancy and inter-gene redundancy by integrating mutual information. Our method employs a dual-phase search strategy: an initial PSO search followed by a local search to accelerate convergence, and a subsequent Pareto front selection to extract the non-dominated gene subsets. Experiments on cancer microarray benchmark datasets demonstrate that our approach significantly enhances feature selection and diagnosis accuracy compared to existing methods. Notably, our approach incorporates a novel dual-evaluation framework and an improved particle representation scheme, which collectively enhance robustness and prevent premature convergence. These innovations ensure a comprehensive and effective gene selection process for cancer diagnosis.

The secure sharing and privacy protection of medical data are of great significance during the development of smart medical care. In order to achieve data sharing among medical institutions, ciphertext-policy attribute-based encryption (CP-ABE), a potential technology, allows users to encrypt data under access policies which are defined on certain attributes of the data consumer, and only allows the data consumer to decrypt those attributes conforming to the access policy. However, some existing CP-ABE schemes still have some shortcomings. For example, the efficiency of encryption and decryption is not high enough, and some cannot support more sufficient and expressive access structures. To solve the above problems, combined with blockchain, this paper presents a CP-ABE scheme with partial policy hiding based on prime order bilinear groups. Extensive experiment and analysis results reveal that the proposed scheme protects the privacy of users and realizes that attribute values are hidden in the access policy.

In geotechnical engineering, the maximum dry density (MDD) stands as an important parameter, denoting the utmost mass of soil achievable per unit volume when compacted to its maximum dry state. Its significance extends to the design of various earthworks like embankments, foundations, and pavements, influencing the soil’s strength, stiffness, and stability. The MDD is contingent on diverse elements like soil type, grain size distribution, moisture content and compaction effort. Generally, heightened compaction effort correlates with an increased MDD, while elevated moisture content corresponds to a reduced MDD. Accurate prediction of the MDD under specific conditions is imperative to uphold the quality and safety standards of earthworks. This research aims to introduce Support Vector Regression (SVR) as a modeling technique for predicting the MDD of soil-stabilizer mixtures. To establish an accurate and comprehensive model that can correlate the stabilized soil’s MDD with attributes of natural soil, consisting linear shrinkage, particle size distribution, plasticity, as well as the type and number of stabilizing additives, three optimization algorithms, namely Artificial Rabbits Optimization (ARO), Manta Ray Foraging Optimization (MRFO), and Improved Manta-Ray Foraging Optimizer (IMRFO), were employed in addition to SVR. Considering the results of evaluative metrics, the SVAR model (combination of SVR and ARO) experienced the highest predictive performance, registering an impressive value of R² in the training phase with 0.9948, as well as the lowest RMSE value of 19.1376.

Accurate indoor localization remains a significant challenge due to the complex nature of indoor environments. This paper proposes a novel method for constructing a radio map (RM) based on Kernel density estimation (KDE) and human trajectories (HT) to enhance indoor localization accuracy. The proposed method utilizes historical HT data in RM construction to capture the spatial variability and complexity of indoor environments, which is crucial for accurate localization. By employing KDE, kernel density maps are generated, identifying high-density regions where additional interpolated fingerprints are strategically placed to improve localization accuracy. In contrast to the conventional method of uniformly placing interpolated points (IPs), the proposed approach better models natural walking patterns and trajectories, thereby enhancing the uniqueness and accuracy of user position identification. Through extensive experiments with various HT patterns, the proposed KDE-RM optimization method consistently outperforms the conventional approach of evenly distributed IPs using Kriging and inverse distance weighting interpolation by up to 36.4%. This demonstrates the effectiveness and potential of the proposed method as a valuable tool for enhancing indoor localization.

The utilization of Wireless Sensor Networks (WSN) in the agricultural field represents a significant stride in the application of Information Technology. Recent advancements in technology have made it possible for sensor networks not only to provide real-time information about soil nutrient levels but also to assist in the automation of various agricultural processes. However, it’s crucial to acknowledge a substantial limitation associated with WSN, namely, energy consumption. Through the analysis of experimental data gathered from diverse soil types and employing sophisticated data analytics, it has been observed that the Nutrient Index exhibits a relatively stable pattern over time. Consequently, predictive neural network techniques can be employed to extract detailed insights from the primary inputs received from WSN. This approach eliminates the need for continuous operation of the WSN throughout the day, contributing to enhanced energy efficiency. To achieve this energy-efficient operation, the NR-MDEC protocol is implemented in conjunction with a coordination algorithm, resulting in a substantial improvement in overall efficiency.

The topic of persuasion in online conversations has social, political and security implications; as a consequence, the problem of predicting persuasive comments in online discussions is receiving increasing attention in the literature. Following recent advancements in graph neural networks, we analyze the impact of conversation structure in predicting persuasive comments in online discussions. We evaluate the performance of artificial intelligence models receiving as input graphs constructed on the top of online conversations sourced from the “Change My View” Reddit channel. We experiment with different graph architectures and compare the performance on graph neural networks, as structure-based models, and dense neural networks as baseline models. Experiments are conducted on two tasks: (1) persuasive comment detection, aiming to predict which comments are persuasive, and (2) influence prediction, aiming to predict which users are persuasive. The experimental results show that the role of the conversation structure in predicting persuasiveness is strongly dependent on its graph representation given as input to the graph neural network. In particular, a graph structure linking only comments belonging to the same speaker in the conversation achieves the best performance in both tasks. This structure outperforms both the baseline model, which does not consider any structural information, and structures linking different speakers’ comments with each other. Specifically, the F1 score of the best performing model is 0.58, which represents an improvement of 5.45% over the baseline model (F1 score of 0.55) and 7.41% over the model linking different speakers’ comments (F1 score of 0.54).

Advanced metering infrastructure (AMI) plays a critical role in the smart grid by integrating metering systems with communication capabilities, especially for the industrial internet of things. However, existing authentication protocols have proven ineffective against quantum computing attacks and are computationally intensive since AMI contains limited computing components, such as smart meters. In this paper, we present a novel, efficient module learning with errors-based authentication and key agreement system for AMI, which we call EMAS. As part of the security measures of EMAS, Kyber Post-Quantum Public Key Encryption, a one-time pad mechanism, and hash functions are used. A formal and informal analysis of the security features is presented, showing that the proposed system is secure and resistant to known attacks. The performance analysis of our proposed EMAS on a B-L475E-IOT01A node equipped with a ARM Cortex M4 microcontroller shows that EMAS is more efficient than existing relevant schemes. About the computation time, EMAS takes 15.693 ms. This result is lower than other existing relevant schemes.

The significance of facial recognition in the era of artificial intelligence lies in its utilization of facial features as a type of biometric characteristic possessing uniqueness and irreversibility. However, exposing these features to attacks, tampering, or unauthorized disclosure poses considerable threats to user privacy and security. A privacy and security solution based on deep learning and encryption algorithms is proposed to tackle this issue. This solution employs the FaceNet deep learning algorithm to extract facial features efficiently. The combination of biometric feature blurriness and cryptographic system precision is achieved, utilizing the CKKS fully homomorphic encryption algorithm for operations in the ciphertext domain of facial recognition. The SM4 algorithm is used to enhance the resilience of facial feature ciphertext against malicious attacks. By leveraging the properties of symmetric ciphers, a balance is achieved between security and computational efficiency. The management of the symmetric key used in the SM4 algorithm is conducted through the employment of the SM9 asymmetric encryption algorithm. Experimental results and analysis demonstrate that the proposed solution enhances the security of data transmission, storage, and comparison without compromising the accuracy and efficiency of facial recognition.

Free Space Optical (FSO) is a wireless communication technology that is distinguished from other communication systems by several advantages including license free of operating spectrum, high data rate, low installation cost, and fast deployment. FSO is employed in many applications including Internet and mobile services links. Nevertheless, FSO link quality is affected by weather conditions including fog, rain, and snow. The main challenge of the FSO channel is the dynamic fluctuating of these weather conditions which degrade the link quality and reduces the data rate. Therefore, the development of robust FSO link topology is a crucial issue to overcome the bad and severe weather conditions. Cluster-based Wireless Sensor Network (WSN) arranges the network into groups called clusters where one Cluster Head (CH) is selected to manage the communication activities inside the group. CHs localization is the main challenge in cluster-based WSN. The key objective of this research is to develop cluster-based WSN that employs the FSO links to interconnect the CHs to each other. Optimal Cluster Head Localization (OCHL) algorithm is developed to optimally determined the locations of CHs so that the network diversity and coverage area of CHs are improved. Genetic Algorithm (GA) technique is used to obtain a near-optimal solution for the proposed fitness function. Simulation results show that the proposed OCHL algorithm improves the network diversity and coverage area of cluster-based WSN. The weighting parameter of the proposed fitness function can be adjusted to control the effects of covered areas, and link diversity in the fitness function. Additionally, increasing the number of CHs leads to improve the covered area and link diversity. Furthermore, with growing of the number of GA iterations, a better solution for the proposed optimization problem is obtained. Moreover, the Bit Error Rate and Signal to Noise Ratio of FSO links are evaluated based on the rain rate, snow rate, fog, transmitted power, transmitter and receiver aperture diameters, FSO communication range, and weighting parameter. The results demonstrate that the normalized covered area in case of using the proposed OCHL algorithm outperforms as compared to NFCA and LEACH algorithms with 12.95 and 8.52% rise, respectively. In addition, the proposed OCHL algorithm enhances the normalized link diversity by 14.15 and 19.21%, compared with NFCA and LEACH algorithms, respectively.

In the competitive landscape of the telecom sector, a Communication Service Provider's success hinges on its ability to offer compelling mobile plans tailored to diverse customer needs. This not only boosts company profits but also enhances metrics like average revenue per user (ARPU), customer lifecycle value, and reduces customer churn. Striking a balance between these objectives presents a formidable task. To address this challenge, we propose a novel approach called Genetic Algorithm Mobile Plan Generation (GA-MPG). The proposed method stands out for its deterministic approach that equally focuses on minimizing customer churn. This is done by providing them with the best-suited plans without making them pay extra for features they would use. The efficient mobile plan generation using GA-MPG is accomplished by the combination of the AdaBoost classifier and the Fuzzy model. The AdaBoost is utilized for feasible mobile plan generation and predicting the optimal solution amongst the various plans. Additionally, a fuzzy model recommends personalized plans based on customers' typical service usage. This also maximizes company profits, contrasting with existing strategies employed by various telecom companies which focus on one of the two problems. The proposed GA-MPG algorithm demonstrated promising results on a prominent US-based telecom dataset encompassing around 7000 customers, with a substantial 44% reduction in customer churn. These findings are based on the simulation results. The algorithm also shows improvements of 13% and 18% in ARPU and company profit, respectively, over a defined period.