Breast cancer is the most common cancer among women and globally affects both genders. The disease arises due to abnormal growth of tissue formed of malignant cells. Early detection of breast cancer is crucial for enhancing the survival rate. Therefore, artificial intelligence has revolutionized healthcare and can serve as a promising tool for early diagnosis. The present study aims to develop a machine-learning model to classify breast cancer and to provide explanations for the model results. This could improve the understanding of the diagnosis and treatment of breast cancer by identifying the most important features of breast cancer tumors and the way they affect the classification task. The best-performing machine-learning model has achieved an accuracy of 97.7% using k-nearest neighbors and a precision of 98.2% based on the Wisconsin breast cancer dataset and an accuracy of 98.6% using the artificial neural network with 94.4% precision based on the Wisconsin diagnostic breast cancer dataset. Hence, this asserts the importance and effectiveness of the proposed approach. The present research explains the model behavior using model-agnostic methods, demonstrating that the bare nuclei feature in the Wisconsin breast cancer dataset and the area’s worst feature Wisconsin diagnostic breast cancer dataset are the most important factors in determining breast cancer malignancy. The work provides extensive insights into the particular characteristics of the diagnosis of breast cancer and suggests possible directions for expected investigation in the future into the fundamental biological mechanisms that underlie the disease’s onset. The findings underline the potential of machine learning to enhance breast cancer diagnosis and therapy planning while emphasizing the importance of interpretability and transparency in artificial intelligence-based healthcare systems.

N ext-generation wireless systems are poised to de-liver unprecedented advancements in data rates, latency, and coverage compared to their predecessors. A key enabler for achieving these objectives is channel coding technology, which is critical in reducing error rates in wireless communications. To meet the demanding requirements of next-generation wireless systems, advanced channel coding techniques such as polar codes, low-density parity check codes, and Turbo codes have been proposed. These innovative coding schemes outperform traditional convolutional and Reed-Solomon codes in terms of error-correction capability and spectral efficiency. This paper presents a comprehensive overview of the essential channel coding technologies employed in next-generation (5G) wireless systems, specifically focusing on fifth-generation systems and beyond. Additionally, it highlights some of the significant research challenges and opportunities in this field, shedding light on the advancements needed to exploit the potential of future wireless communication networks fully. Compared to other survey papers in the field, our objective is to provide a more thorough summary of the most relevant channel coding techniques. This enables researchers and application developers to quickly learn the 5G channel codes. Additionally, we provide a comparison between the channel codes.

This work presents an extensive dataset comprising images meticulously obtained from diverse geographic locations within Iraq, depicting both healthy and infected fig leaves affected by Ficus leafworm. This particular pest poses a significant threat to economic interests, as its infestations often lead to the defoliation of trees, resulting in reduced fruit production. The dataset comprises two distinct classes: infected and healthy, with the acquisition of images executed with precision during the fruiting season, employing state-of-the-art high-resolution equipment, as detailed in the specifications table. In total, the dataset encompasses a substantial 2,321 images, with 1,350 representing infected leaves and 971 depicting healthy ones. The images were acquired through a random sampling approach, ensuring a harmonious blend of balance and diversity across data emanating from distinct fig trees. The proposed dataset carries substantial potential for impact and utility, featuring essential attributes such as the binary classification of infected and healthy leaves. The presented dataset holds the potential to be a valuable resource for the pest control industry within the domains of agriculture and food production.

The use of deep learning techniques with matrix multiplication has transformed several fields, leading to notable breakthroughs and innovative uses. In this paper, we summarize the state-of-the-art and explain the many applications, approaches, and advancements resulting from the combination of matrix multiplication with deep learning. Examining a wide range of articles released since 2019, this evaluation focuses on the creative methods, strategies, and improvements that scholars in many domains have used. The review encapsulates significant strides in matrix multiplication optimization through deep learning methodologies, unveiling a spectrum of innovative approaches and their consequential impacts. These advancements span dynamic precision neural network inference on silicon photonic processors, adaptable templates for matrix multiplication across various architectures, and coherent analog computing engines for optical matrix multiplication. Furthermore, the exploration extends to unconventional accelerators like AI Engine tiles and low-power multi-layer perceptron accelerators, showcasing remarkable efficiency gains. The integration of machine learning frameworks, such as WISE, into sparse matrix operations further underscores the potential for substantial computational efficiency improvements. Collectively, these findings underscore the transformative potential of deep learning-driven optimization techniques in revolutionizing matrix multiplication performance, offering pathways to enhanced computational efficiency and energy savings across diverse application domains.

Recent research has revealed that using machine learning systems for the analysis of genetic data could reliably detect Alzheimer’s disease. The interpretability of these models, however, has been a challenge, as they frequently provided little insight into the features that contribute to their predictions. Explainable machine learning has been presented as a solution to this problem since it enables the identification of significant attributes and gives a clearer method of making predictions. In this study, Genome-Wide Association Studies were used to recognize genetic variants associated with Alzheimer’s disease, utilizing the Alzheimer’s Disease Neuroimaging Initiative dataset and quality control methods to ensure the validity and reliability of the findings. The results indicate strong connections between certain genetic variations and Alzheimer’s disease, highlighting the potential of Genome-Wide Association Studies as a valuable tool for identifying and predicting this disease. After studying and analyzing the genetic data, machine learning algorithms are utilized to train a model to detect Alzheimer. The Support Vector Machine achieved 89% accuracy as the best-performing model. Explainable machine learning has the potential to increase the accuracy and interpretability of Alzheimer’s disease detection models, giving significant insights for both academics and physicians. The explanation of the support vector machine model reveals that rs4821510 is the most important SNP in detecting AD. On top of that, the SHAP method shows that rs429358 is an indication for Alzheimer’s disease and rs4821510 presents in the healthy ones. These findings suggest that explainable machine learning can play an important role in accurately detecting Alzheimer’s disease and identifying critical genetic markers associated with the disease.

Losses in the electrical power transmission and distribution systems are considered two of the most critical challenges in power grids. Reducing the related losses plays a significant role in increasing system efficiency in addition to diminishing costs. Therefore, optimum power transfer as well as finding a convenient route, are essential factors in electrical grids. This paper intends to substantially reduce the transmission/distribution-related losses by finding the shortest and most optimal path between the renewable energy power plant (producer) and the substations/consumers. A genetic algorithm (GA) is proposed for optimal routing to increase the system’s reliability and minimize the losses of the entire network. In this work, by presenting a coding with chromosomes of variable length and considering the construction costs and the power transmission line/path as the fitness function, the appropriate route is obtained. The efficiency of the proposed method is compared with Dijkstra’s algorithm, one of the conventional graph search approaches. The ant colony optimization (ACO) algorithm and a reinforcement learning algorithm, namely the Q-learning model, are employed to further explore the optimization efficiency of the proposed renewable energy-based transmission system. The simulation results demonstrate that the proposed models accurately determine the optimal pathway within an excellent time.

The use of massive multiple-input multiple-output (mMIMO) technology is essential for the fifth-generation (5G) and sixth-generation (6G) networks. However, the computational complexity of detection techniques and approximation methods can be high due to matrix inversion. Deep learning (DL) has been proposed as a tool to improve the efficiency of massive MIMO systems. This study proposes a hybrid-based low-complexity detector employing deep learning and approximate matrix inversion. The Richardson method and multi-scale multi-skip connection network (MMNet) form the presented hybrid detection framework. The output of the first iteration of approximate matrix inversion methods is fed into the MMNet algorithm in order to obtain superior performance. The results are compared with the MMSE-based and conventional MMNet-based detectors to determine/benchmark the performance. The simulation results and benchmarks with an MMSE-based and conventional MMNet-based detectors further designate that employing the proposed model significantly enhances the detection performance.

This paper performs speech emotion recognition on short voice messages lasting less than three seconds, using one-dimensional convolutional neural networks. The Ravee dataset, voiced by professional actors, is exploited. The proposed convolutional neural network architecture for the speech emotion recognition system aims to improve accuracy and reduce the total processing cost of the speech emotion recognition model. Moreover, Mel-frequency cepstral coefficients are used as the main features for recognition purposes. Additionally, overfitting problems are avoided by utilizing data augmentation techniques and feature extraction algorithms, which enhance testing ac-curacy by increasing the number of training samples. Various simulations are conducted, through which it is observed that the proposed model provides recognition accuracy of up to 83%.

Nowadays, robotic applications exist in various fields, including medical, industrial, and educational. The critical aspect of most of these applications is robot movement, where an efficient path-planning algorithm is required in order to guarantee a safe and cost-effective movement. The main goal of the path planning technique is to find the shortest possible path to the destination while avoiding the obstacles on the route. This study proposes a framework employing swarm intelligence optimization techniques based on an improved genetic algorithm and particle swarm optimization to obtain the optimum trajectory. The simulations are conducted using MATLAB R2022b. It is observed that the proposed particle swarm optimization achieves better accuracy of up to 99.5% and faster convergence time when compared with the genetic algorithm that attains 74.6% accuracy. The proposed optimized path planning algorithm is considerably advantageous, especially in realistic applications such as rescue robots and item delivery.

Blind source separation is a challenging problem in signal processing, involving the separation of mixed signals into their individual sources. This paper introduces a novel approach based on multiuser kurtosis to address the blind source separation problem. The proposed technique utilizes the kurtosis of the signals to estimate both the source signals and the mixing matrix. By leveraging the higher-order statistical properties of the signals, particularly their fourth-order statistics, the mixing matrix is estimated and used to separate the original sources. To enhance the separation performance, a multiuser extension of the kurtosis-based approach is introduced, enabling simultaneous separation and retrieval of multiple sources. This extension employs a joint diagonalization approach to estimate the mixing matrix and perform source separation. The performance of the proposed model is evaluated on synthetic and real-world datasets, and compared against other state-of-the-art blind source separation techniques. The experimental results demonstrate that the multiuser kurtosis-based algorithm outperforms existing methods in terms of separation accuracy and computational efficiency. Furthermore, the algorithm exhibits robustness to noise and can handle non-linear mixing models. Additionally, a genetic algorithm is employed in this study to further enhance the separation/estimation performance of the multiuser kurtosis-based blind source separation. The potential of the proposed model in speech and image processing applications is demonstrated, showing competitive performance compared to existing techniques. The simulation results confirm the promising nature of the proposed genetic algorithm-optimized multiuser kurtosis-based method for solving blind source separation problems in various signal and image processing applications.

People's lives always necessitate prioritizing having the best possible and accessible water quality. Since pollution sources continuously expand, water quality monitoring systems have become essential. However, such monitoring systems remain too expensive. Fortunately, with the recent growth of the Internet of Things (IoT) technology and the ongoing development of smart monitoring systems, a low-cost smart water quality monitoring system can be implemented. This paper presents a novel methodology for water quality monitoring systems, allowing smart acquisition of water data, real-time monitoring, and the maintenance of the optimal water quality required for the intended audience. This work provides the necessary monitoring options in a reliable and cost-efficient manner. The system design consisting of hardware and software architectures is presented and investigated.

ChatGPT is an extensive language model under the umbrella of generative artificial intelligence that produces answers from data and images curated from online resources. Despite the capability to produce accurate responses, but requires verification; the responses are based on statistical patterns rather than true comprehension, i.e., it does not have consciousness and does not understand the questions from the perspective of human comprehension. The ability of ChatGPT to understand and react to questions in a humanistic way has garnered a lot of public and scientific interest over the past year. This study analyzes responses of ChatGPT to 100 questions on epilepsy in order to assess the validity of the tool in this field. Besides, this work sheds light on the advantages and disadvantages of the approach in this particular topic by analyzing responses of ChatGPT to queries on epilepsy. The study evaluates the model performance by looking at the completeness, correctness, and relevancy of responses. The findings in this paper indicate that ChatGPT has limits because of its training data and design structure, even though it could give insightful and appropriate answers to inquiries about epilepsy. It is concluded that ChatGPT can be an advantageous tool for medical professionals working on the subject of epilepsy. Nonetheless, it should be noted that ChatGPT should be utilized cautiously and in conjunction with various information sources, like clinical practice guidelines and peer-reviewed studies.

Skin cancer is the most common type of cancer worldwide, affecting a large population recently. To date, various machine learning techniques exploiting skin images have been applied directly to skin cancer classification, showing promising results in improving diagnostic accuracy. This study aims to develop a machine learningbased model capable of accurately classifying skin cancer by utilizing extracted features from preprocessed images in the publicly available PH2 dataset. Preprocessed features are known to provide more significant information than raw image data, as they capture specific characteristics of the images that are relevant to the classification task. The proposed model of this study can identify the most pertinent information in the images more accurately, thereby improving the performance and interpretability of the machine learning classification. Our simulation results illustrate that employing XG-boost yields an accuracy of 94% and an area under the curve value of 0.9947, further indicating that the proposed technique effectively distinguishes between non-melanoma and melanoma skin cancer. Explainable artificial intelligence provides some explanations by leveraging modelagnostic methods such as partial dependence plot, permutation importance, and SHAP. Moreover, the explainable artificial intelligence results show that asymmetry and pigment network features are the most important feature in the classification of skin cancer. These specific characteristics emerge as the most influential factors in distinguishing between different types of skin cancer.

Molecular communication (MC) is a modern communication paradigm inspired by biological mechanisms and systems. Due to the short range of molecular diffusion, MC systems necessitate a multi-hop diffusion-based network to transmit information. Finding the optimal routing path is one of the most critical challenges in MC. The main goal is to transfer information through the diffusion of molecules within an optimal state by detecting the shortest route and the proper relays. In this paper, finding the optimal routing path using a genetic algorithm (GA) is investigated in order to find the shortest and the most energy-efficient path. Our model intelligently plans the optimum trajectory between the transmitter (TX) and the receiver (RX) by identifying the appropriate relays both locally and globally. Our GA implementation uses a variable-length chromosome encoding to obtain the optimal path by selecting an appropriate fitness function. We also examine and compare the performance of the proposed algorithm with Dijkstra’s algorithm (DA), which is one of the deterministic algorithms. Finally, various simulations for different sizes of MC networks are performed to verify the accuracy of the proposed method. Our simulation results demonstrate that the presented GA offers an accurate routing path within an excellent time, even in large-sized environments.

Almost all signals existing in the universe experience varying degrees of noise interference. Specifically, audio signals necessitate efficient noise cancellation for most hearing devices to comfort the user. Various filtering techniques are employed in order to apply efficient noise cancellation, empowering the system to enhance the signal-to-noise ratio. Currently, adaptive filters are preferred to other types of filters to approach higher efficiency. This study presents and examines four adaptive filter algorithms, including least-mean-square, normalized least-mean-square, recursive-least-square, and Wiener filter. The selected models are simulated, benchmarked, and contrasted in some characteristics of the performance. The presented filters are applied to four different experiments/environments to further examine their functionality. All of that is performed utilizing different step sizes to monitor two compromised result parameters: performance and execution time. Eventually, the best adaptive filter possessing the optimal parameters and step size is acquired for electrocardiogram signals enabling physicians and health professionals to deal with electrocardiogram signals efficiently, empowering them to accurately and quickly diagnose any sign of heart problems. Simulation results further designate the superiority of the presented models.

Medication non-adherence is a prevalent concern, particularly among individuals managing chronic illnesses who rely on consistent pill consumption. This study addresses forgetfulness and non-compliance in medication intake by proposing a system that ensures accurate administration of prescribed medications at designated times. This paper investigates medication adherence challenges, primarily focusing on chronic condition management. Leveraging mobile phones, our innovative approach aims to mitigate these challenges. This paper proposes a system that delivers timely reminders to patients via mobile devices, fostering responsibility toward adhering to medication regimens. Central to the proposed solution is a patient-to-hospital communication framework, enabling caregivers to curate medication schedules. Caregivers have control over medications, timings, and dosages. This empowers short-term and long-term medication users to monitor regimens, alleviating concerns of omissions or deviations. Implications of the proposed framework are far-reaching. Consistent medication adherence can enhance therapeutic interventions, potentially reducing morbidity and mortality. The presented technology-healthcare convergence underscores the positive impact of technological interventions in medical contexts.

This work provides a general review of Adaptive Modulation and Coding (AMC) techniques that optimize the utilization of available resources and parameters in wireless communication systems. AMC or link adaptation empowers communication systems to employ appropriate modulation and coding schemes in order to enhance link spectrum utilization over time-varying channels. In addition, the proper selection of Modulation and Coding Schemes (MCS) diminishes the Bit Error Rate (BER) of systems and reduces the required transmit power through which the interference to other users is minimized. This paper conducts a state-of-the-art review of the existing AMC technologies. The reviews provided in this paper shed light on the trends of research in AMC and bring focus to related techniques. First, the existing constraints and the related phenomena in advanced wireless communication, including channel capacity, path loss, fading, and interference, are thoroughly investigated to understand further and improve the AMC systems. Moreover, several scenarios of utilizing the AMC methods are provided in order to demonstrate the extent of the effectiveness and importance of accurate AMC. Lastly, the challenges, the likely problems, the research gaps, and the potential avenues for future work are declared.

In various signal processing applications, such as audio signal recovery, the extraction of desired signals from a mixture of other signals is a crucial task. To achieve superior performance and efficiency in separator systems, extensive research has been conducted. Blind source separation emerges as a relevant technique to address the challenge of separating and reconstructing unknown signals when only observations of their mixtures are available to end-users. Blind source separation involves retrieving a set of independent source signals mixed by an unknown and potentially destructive combining system. Notably, the separation process in blind source separation frameworks solely relies on observing the mixed sources without prior knowledge of the mixing algorithm or the source signal characteristics. The significance of blind source separation has garnered substantial attention, and its numerous applications have been demonstrated, which serves as the primary motivation for conducting this comprehensive study. This paper presents a systematic literature survey of blind source separation, encompassing existing methods, approaches, and applications, with a particular focus on artificial intelligence-based frameworks. Through a thorough review and examination, this work sheds light on the diverse techniques utilized in blind source separation and their performance in real-world scenarios. The study identifies research gaps in the current literature, highlighting areas that warrant further investigation and improvement. Moreover, potential avenues for future research are outlined to contribute to the ongoing development of blind source separation techniques.

In the contemporary era, blind source separation has emerged as a highly appealing and significant research topic within the field of signal processing. The imperative for the integration of blind source separation techniques within the context of beyond fifth-generation and sixth-generation networks arises from the increasing demand for reliable and efficient communication systems that can effectively handle the challenges posed by high-density networks, dynamic interference environments, and the coexistence of diverse signal sources, thereby enabling enhanced signal extraction and separation for improved system performance. Particularly, audio processing presents a critical domain where the challenge lies in effectively handling files containing a mixture of human speech, silence, and music. Addressing this challenge, speech separation systems can be regarded as a specialized form of human speech recognition or audio signal classification systems that are leveraged to separate, identify, or delineate segments of audio signals encompassing human speech. In various applications such as volume reduction, quality enhancement, detection, and identification, the need arises to separate human speech by eliminating silence, music, or environmental noise from the audio signals. Consequently, the development of robust methods for accurate and efficient speech separation holds paramount importance in optimizing audio signal processing tasks. This study proposes a novel three-way neural network architecture that incorporates transfer learning, a pre-trained dual-path recurrent neural network, and a transformer. In addition to learning the time series associated with audio signals, this network possesses the unique capability of direct context-awareness for modeling the speech sequence within the transformer framework. A comprehensive array of simulations is meticulously conducted to evaluate the performance of the proposed model, which is benchmarked with seven prominent state-of-the-art deep learning-based architectures. The results obtained from these evaluations demonstrate notable advancements in multiple objective metrics. Specifically, our proposed solution showcases an average improvement of 4.60% in terms of short-time objective intelligibility, 14.84% in source-to-distortion ratio, and 9.87% in scale-invariant signal-to-noise ratio. These extraordinary advancements surpass those achieved by the nearest rival, namely the dual-path recurrent neural network time-domain audio separation network, firmly establishing the superiority of our proposed model’s performance.

As the number of users within a cellular system increase, so does the need for a higher quality of service and performance. Massive multiple-input multiple-output (MIMO) is a staple technology in the implementation of fifth-generation (5G) cellular networks. The technology leverages multiple antennas at the base station and within user devices to increase spectral efficiency, link reliability, and range. With massive MIMO being a significant area of research, several data detection techniques exist. Classical methods leveraging linear detection algorithms and linear approximate matrix inversions exist. However, as of recently, detection algorithms utilizing deep learning (DL) have been proposed. The detection problem requires techniques that are both robust and provide near-optimal performance at the expense of minimal complexity within different channel scenarios. DL utilizes machine learning to train a detection algorithm providing comparable performance to classic techniques with the advantage of lower complexity. This paper offers a hybrid detection algorithm consisting of a linear approximate matrix inversion step followed by a DL algorithm. The approximated message vector calculated in the first step is the initial iterate fed to the DL algorithm, which continuously improves the detection accuracy. Various simulations are conducted, demonstrating the significant superiority of the proposed framework. The study is concluded with an analysis of the complexity of the hybrid algorithm in addition to a discussion of the model's performance.