Real-world Applications Research Articles

Reliable Autism Spectrum Disorder Diagnosis for Pediatrics Using Machine Learning and Explainable AI

Background: As the demand for early and accurate diagnosis of autism spectrum disorder (ASD) increases, the integration of machine learning (ML) and explainable artificial intelligence (XAI) is emerging as a critical advancement that promises to revolutionize intervention strategies by improving both accuracy and transparency. Methods: This paper presents a method that combines XAI techniques with a rigorous data-preprocessing pipeline to improve the accuracy and interpretability of ML-based diagnostic tools. Our preprocessing pipeline included outlier removal, missing data handling, and selecting pertinent features based on clinical expert advice. Using R and the caret package (version 6.0.94), we developed and compared several ML algorithms, validated using 10-fold cross-validation and optimized by grid search hyperparameter tuning. XAI techniques were employed to improve model transparency, offering insights into how features contribute to predictions, thereby enhancing clinician trust. Results: Rigorous data-preprocessing improved the models’ generalizability and real-world applicability across diverse clinical datasets, ensuring a robust performance. Neural networks and extreme gradient boosting models achieved the best performance in terms of accuracy, precision, and recall. XAI techniques demonstrated that behavioral features significantly influenced model predictions, leading to greater interpretability. Conclusions: This study successfully developed highly precise and interpretable ML models for ASD diagnosis, connecting advanced ML methods with practical clinical application and supporting the adoption of AI-driven diagnostic tools by healthcare professionals. This study’s findings contribute to personalized intervention strategies and early diagnostic practices, ultimately improving outcomes and quality of life for individuals with ASD.

GTBTL-IoT: An Approach of Curtailing Task Offloading Time for Improved Responsiveness in IoT-MEC Model

INTRODUCTION: The Internet of Things (IoT) has transformed daily life by interconnecting digital devices via integrated sensors, software, and connectivity. Although IoT devices excel at real-time data collection and decision-making, their performance on complex tasks is hindered by limited power, resources, and time. To address this, IoT is often combined with cloud computing (CC) to meet time-sensitive demands. However, the distance between IoT devices and cloud servers can result in latency issues. OBJECTIVES: To mitigate latency challenges, Mobile Edge Computing (MEC) is integrated with IoT. MEC offers cloud-like services through servers located near network edges and IoT devices, enhancing device responsiveness by reducing transmission and processing latency. This study aims to develop a solution to optimize task offloading in IoT-MEC environments, addressing challenges like latency, uneven workloads, and network congestion. METHODS: This research introduces the Game Theory-Based Task Latency (GTBTL-IoT) algorithm, a two-way task offloading approach employing Game Matching Theory and Data Partitioning Theory. Initially, the algorithm matches IoT devices with the nearest MEC server using game-matching theory. Subsequently, it splits the entire task into two halves and allocates them to both local and MEC servers for parallel computation, optimizing resource usage and workload balance. RESULTS: GTBTL-IoT outperforms existing algorithms, such as the Delay-Aware Online Workload Allocation (DAOWA) Algorithm, Fuzzy Algorithm (FA), and Dynamic Task Scheduling (DTS), by an average of 143.75 ms with a 5.5 s system deadline. Additionally, it significantly reduces task transmission, computation latency, and overall job offloading time by 59%. Evaluated in an ENIGMA-based simulation environment, GTBTL-IoT demonstrates its ability to compute requests in real-time with optimal resource usage, ensuring efficient and balanced task execution in the IoT-MEC paradigm. CONCLUSION: The Game Theory-Based Task Latency (GTBTL-IoT) algorithm presents a novel approach to optimize task offloading in IoT-MEC environments. By leveraging Game Matching Theory and Data Partitioning Theory, GTBTL-IoT effectively reduces latency, balances workloads, and optimizes resource usage. The algorithm's superior performance compared to existing methods underscores its potential to enhance the responsiveness and efficiency of IoT devices in real-world applications, ensuring seamless task execution in IoT-MEC systems.

The Road to Explainable Graph Neural Networks

GOAL: Graph Neural Networks (GNNs) are being increasingly adopted in various real-world applications, including drug and material discovery [16, 17, 27], recommendation engines [29], congestion modeling on road networks [5], and weather forecasting [11]. However, similar to other deeplearning models, GNNs are considered black boxes due to their limited capacity to provide explanations for their predictions. This lack of interpretability poses a significant barrier to their adoption in critical domains such as healthcare, finance, and law enforcement, where transparency and trustworthiness are essential for decision-making processes. In these pivotal domains, understanding the rationale behind model predictions is crucial not only for compliance with interpretability requirements but also for identifying potential vulnerabilities and gaining insights to refine the model further. How do we make GNNs interpretable? This is a central motivating question driving my research pursuits.

Students’ Perceptions on Chemistry Education Programme: A Qualitative Inquiry in Faculty of Education

This article examines the efficacy of the Chemistry Education program offered by the Department of Science and Environment Education, Central Department of Education, Tribhuvan University. It also explores students' perceptions and experiences regarding the program, focusing on its strengths and challenges. Using purposive sampling, ten students enrolled in the Science Education program were selected for data collection through face-to-face interviews and conversations. Verbatim and conversational analyses were employed to analyze the data. The findings indicate that the culture of teaching and learning has shifted from traditional, teacher-centered approaches to student-centered pedagogies, with a strong emphasis on research-based and project-oriented learning. Students appreciated the comprehensive coverage of theoretical concepts in the Chemistry Education program. However, they expressed concerns about the limited opportunities for practical laboratory work, which they felt hindered their ability to fully grasp and apply the concepts learned in class. Additionally, the study found that while the program integrates theoretical constructs with real-world applications through experiential learning modalities, doubts persist regarding its effectiveness in cultivating skills essential for long-term professional success. It is strongly recommended to elevate the Chemistry Education program by integrating more dynamic and interactive pedagogical approaches to more effectively nurture the skills critical for sustained professional success.

Bias and Its Consequences : A Study of Machine Learning Performance

This paper addresses the concern about bias affecting the results of machine learning models. For this purpose, it uses the Adult Income dataset from OpenML for income classification. The conditions for bias are induced by underrepresenting people that earn <= $50K in training data, thus checking the behavior of different models when encountering such a skewed distribution. Key metrics, namely accuracy and specificity (True Negative Rate), were analyzed for unbiased and biased training scenarios. The results show that Naive Bayes and Random Forest models were resistant to bias, but others, including SVM and Logistic Regression, suffered major performance drops. This study throws light on the robustness of different classifiers when exposed to biased data, requiring further bias mitigation strategies in real-world applications. This paper actually examines critically how bias in training data can significantly affect the performance of prediction, fairness, and model selection in income classification tasks.

Enhancing stability and explainability in reinforcement learning with machine learning

Abstract. In the field of reinforcement learning, training agents using machine learning algorithms to learn and perform tasks in complex environments has become a prevalent approach. However, reinforcement learning faces challenges such as training instability and decision opacity, which limit its feasibility in real-world applications. To solve the problems of stability and transparency in reinforcement learning, this project will use advanced algorithms like Proximal Policy Optimization (PPO), Q-DAGGER, and Gradient Boosting Decision Trees to set up reinforcement learning agents in the OpenAI Gymnasium environment. Specifically, the study selected the Atari game Breakout as the testbed, enhancing training efficiency and game performance by refining reward structures and decision-making processes, and integrating interpretable models to provide explanations for agent decisions. This study has successfully developed robust reinforcement learning agents that excel in complex environments. By employing advanced algorithms like PPO, Q-DAGGER, and Gradient Boosting Decision Trees, the study has addressed issues of training instability, and improved game performance through optimized reward structures and decision processes. Additionally, by integrating interpretable models, the study has provided insights into the learned strategies of the agents, thereby enhancing decision transparency. These findings provide crucial support for the broader application of reinforcement learning in real-world scenarios and offer valuable insights for tackling other complex tasks.

Event-Driven Data Pipelines : A Cloud-Based Approach to Real-Time Data Processing

This article comprehensively analyzes event-driven data pipelines in cloud computing environments, examining their architecture, implementation considerations, and real-world applications. The article explores the fundamental components of event-driven systems, from event sources and message brokers to processing layers, while evaluating their performance characteristics and reliability mechanisms. Through detailed analysis of system architectures, we investigate the integration of various cloud services and their role in enabling scalable, real-time data processing. The article demonstrates how modern event-driven architectures achieve sub-millisecond processing times and handle millions of events per second while maintaining system resilience. Our findings reveal significant improvements in operational efficiency across various industries, including financial services, marketing automation, and IoT monitoring solutions. The article also addresses critical implementation aspects, including security frameworks, integration protocols, and best practices for system optimization. These insights provide valuable guidance for organizations leveraging event-driven architectures in their digital transformation initiatives. The article concludes that event-driven data pipelines represent a crucial advancement in data processing technology, offering unprecedented capabilities in handling real-time data streams while maintaining scalability, reliability, and cost-effectiveness.

External Validation of Accelerometry-Based Mechanical Loading Prediction Equations

Accurately predicting physical activity-associated mechanical loading is crucial for developing and monitoring exercise interventions that improve bone health. While accelerometer-based prediction equations offer a promising solution, their external validity across different populations and activity contexts remains unclear. This study aimed to validate existing mechanical loading prediction equations by applying them to a sample and testing conditions distinct from the original validation studies. A convenience sample of 49 adults performed walking, running, and jumping activities on a force plate while wearing accelerometers at their hip, lower back, and ankle. Peak ground reaction force (pGRF) and peak loading rate (pLR) predictions were assessed for accuracy. Substantial variability in prediction accuracy was found, with pLR showing the highest errors. These findings highlight the need to improve prediction models to account for individual biomechanical differences, sensor placement, and high-impact activities. Such refinements are essential for ensuring the models’ reliability in real-world applications, particularly in clinical and biomechanical research contexts, where accurate assessments of mechanical loading are critical for designing rehabilitation programs, injury prevention strategies, and optimizing bone health interventions.

Implementation of Distributed Machine Learning in Cloud Calculation: Evidence from Consumer Behavior Prediction

Abstract. In the field of machine learning, distributed machine learning (DML) has gained massive popularity in research and appeared in a wide range of applications including transportation, medicine, and business. Distributed machine learning showed huge potential in processing large amounts of data while keeping the data private. In this paper, a distributed implementation of a deep neural network (DNN) with a large dataset from a telecom company, aiming to predict consumer behavior based on usage data is carried out. Various datasets with distinct heterogeneity settings were applied to the network, simulating a real-world cloud application scenario. Comparisons between both the centralized model and the distributed models were made to analyze the impact on model performance with device heterogeneity and data heterogeneity. These result indicate that under certain circumstances, DML models showed better accuracy and lower loss compared to the centralized model. The paper provided some insights and made some assumptions on how the data and the parameters affect the model performance.

Towards Pragmatic Temporal Alignment in Stateful Generative AI Systems: A Configurable Approach

Temporal alignment in stateful generative artificial intelligence (AI) systems remains an underexplored area, particularly beyond goal-driven approaches in planning. Stateful refers to maintaining a persistent memory or ``state'' across runs or sessions. This helps with referencing past information to make system outputs more contextual and relevant. This position paper proposes a position framework for temporal alignment with configurable toggles. We present five alignment mechanisms: knowledge graph path-based, neural score-based, vector similarity-based, and sequential process-guided alignment. By offering these interchangeable approaches, we aim to provide a flexible solution adaptable to complex and real-world application scenarios. This paper discusses the potential benefits and challenges of each alignment method and positions the importance of a configurable system in advancing progress in stateful generative AI systems.

A Review of Facial Generation Technology Research

Abstract. Face generation, as a cutting-edge generative technology, has made significant strides in various image synthesis tasks, including facial inpainting, text-to-image translation, and video-based facial animation. Generating realistic and diverse human faces is a critical task in computer vision, with a wide array of real-world applications. Given the remarkable success of face generation models, there has been a growing interest in leveraging advanced techniques to further improve the quality, diversity, and control of generated faces. This paper will provide a comprehensive overview of the state-of-the-art face generation techniques. Specifically, the paper reviews the key approaches in this field, including Generative Adversarial Networks (GANs), Vector Quantized methods, and the more recent Diffusion Models, each of which has contributed significantly to the advancement of face generation. The paper discusses how these models have evolved to handle the complexity of face generation, from capturing subtle facial details to enabling fine-grained control over facial attributes. The paper also explores the major applications of face generation technologies, with a particular emphasis on their use in entertainment, virtual reality, and security. Finally, the paper identifies promising directions for future research in face generation, such as improving the interpretability of models, addressing ethical concerns, and enhancing the ability to generate faces that are both highly realistic and diverse.

CNN-Based Distracted Driver Recognition: Enhancing Accuracy Through Optimized Training Techniques

Abstract. Distracted driving is a significant contributor to vehicle accidents worldwide, underscoring the need for an effective recognition model that can identify risky driver behaviors in real-time and provide timely alerts. This study proposes a Convolutional Neural Network (CNN)-based model designed to accurately detect distracted drivers. To optimize the model's performance, this paper implemented several strategic enhancements. During the preprocessing phase, 90% of the dataset was allocated for training, ensuring a comprehensive learning process, while the remaining 10% was used for validation to assess the model's accuracy. The Adam optimizer was chosen for its ability to dynamically adjust the learning rate, facilitating faster convergence to an optimal solution. Additionally, the Cross-Entropy loss function was employed to amplify errors during training, driving the model to correct inaccuracies more effectively. The model was trained over 25 epochs, resulting in an accuracy rate of nearly 80%. This level of performance demonstrates the models viability for real-world applications, where it can play a critical role in reducing accidents caused by distracted driving. Future research may focus on further refining the model by exploring advanced loss functions, optimizers, and CNN architectures, as well as incorporating more sophisticated data preprocessing techniques.

Analysis of Multivariable Sensor Responses to Multi-Analyte Gas Samples in the Presence of Interferents and Humidity.

This work presents an adaptive sensor signal-processing approach to enable quantification, using a single gas sensor or a small sensor array, of multianalyte mixtures of aromatic hydrocarbons in the presence of various interferents and humidity for environmental-monitoring applications. Dynamic sensor responses are analyzed by extracting multivariable sensing parameters to provide necessary sensitivity and selectivity. This is achieved by integrating the Levenberg-Marquardt-modified, exponentially weighted, recursive-least-squares-estimation (LM-modified EW-RLSE) algorithm and principal-component analysis (PCA). Achieving measured detection limits as low as 3 μg/L (≤1 ppm by volume) for 6 target analytes, the system exhibits excellent PCA cluster separation for all analytes in the mixtures, with reliable identification and accurate quantification, even in the presence of various interferents. Concentration errors of approximately ±5% are obtained for mixtures containing up to 6 BTEX compounds (including chemical isomers) and up to 4 interferents. Additionally, the study investigates the impact of humidity on the polymer/plasticizer-coated shear-horizontal surface acoustic wave (SH-SAW) sensors, demonstrating accurate concentration estimation in a relative humidity range from dry nitrogen to 65%. This sensing-and-multivariate-signal-processing approach is a promising candidate for reliable environmental monitoring in real-world applications.

The Self-Improving Customer Service System: How AI Learns from Every Interaction

This article explores the transformative impact of self-improving AI systems in customer service, examining the parallels between AI and human learning processes, the mechanics of AI self-improvement, and the advantages of AI's learning at scale. It delves into real-world applications and impacts across various industries, highlighting improvements in personalization, efficiency, accuracy, customer satisfaction, and cost reduction. The article also addresses ethical considerations such as data privacy, transparency, and bias mitigation, while offering a future outlook on advanced AI capabilities in emotional intelligence, predictive service, and seamless human-AI collaboration. Through a comprehensive analysis of current technologies and future possibilities, this study provides insights into how self-improving AI is revolutionizing customer service, promising more efficient, personalized, and satisfying experiences for customers.

Revisiting streaming anomaly detection: benchmark and evaluation

Anomaly detection in streaming data is an important task for many real-world applications, such as network security, fraud detection, and system monitoring. However, streaming data often exhibit concept drift, which means that the data distribution changes over time. This poses a significant challenge for many anomaly detection algorithms, as they need to adapt to the evolving data to maintain high detection accuracy. Existing streaming anomaly detection algorithms lack a unified evaluation framework that validly assesses their performance and robustness under different types of concept drifts and anomalies. In this paper, we conduct a systematic technical review of the state-of-the-art methods for anomaly detection in streaming data. We propose a new data generator, called SCAR (Streaming data generator with Customizable Anomalies and concept dRifts), that can synthesize streaming data based on synthetic and real-world datasets from different domains. Furthermore, we adapt four static anomaly detection models to the streaming setting using a generic reconstruction strategy as baselines, and then compare them systematically with 9 existing streaming anomaly detection algorithms on 76 synthesized datasets that have various types of anomalies and concept drifts. The challenges and future research directions for anomaly detection in streaming data are also presented.

Situating Empathy in HCI/CSCW: A Scoping Review

Empathy is considered a crucial construct within HCI and CSCW, yet our understanding of this complex concept remains fragmented and lacks consensus in existing research. In this scoping review of 121 articles from the ACM Digital Library, we synthesize the diverse perspectives on empathy and scrutinize its current conceptualization and operationalization. In particular, we examine the various interpretations and definitions of empathy, its applications, and the methodologies, findings, and trends in the field. Our analysis reveals a lack of consensus on the definitions and theoretical underpinnings of empathy, with interpretations ranging from understanding the experiences of others to an affective response to the other's situation. We observed that despite the variety of methods used to gauge empathy, the predominant approach remains self-assessed instruments, highlighting the lack of novel and rigorously established and validated measures and methods to capture the multifaceted manifestations of empathy. Furthermore, our analysis shows that previous studies have used a variety of approaches to elicit empathy, such as experiential methods and situational awareness. These approaches have demonstrated that shared stressful experiences promote community support and relief, while situational awareness promotes empathy through increased helping behavior. Finally, we discuss a) the potential and drawbacks of leveraging empathy to shape interactions and guide design practices, b) the need to find a balance between the collective focus of empathy and the (existing and dominant) focus on the individual, and c) the careful testing of empathic designs and technologies with real-world applications.

Integrating spoken instructions into flight trajectory prediction to optimize automation in air traffic control

The booming air transportation industry inevitably burdens air traffic controllers’ workload, causing unexpected human factor-related incidents. Current air traffic control systems fail to consider spoken instructions for traffic prediction, bringing significant challenges in detecting human errors during real-time traffic operations. Here, we present an automation paradigm integrating controlling intent into the information processing loop through the spoken instruction-aware flight trajectory prediction framework. A 3-stage progressive multi-modal learning paradigm is proposed to address the modality gap between the trajectory and spoken instructions, as well as minimize the data requirements. Experiments on a real-world dataset show the proposed framework achieves flight trajectory prediction with high predictability and timeliness, obtaining over 20% relative reduction in mean deviation error. Moreover, the generalizability of the proposed framework is also confirmed by various model architectures. The proposed framework can formulate full-automated information processing in real-world air traffic applications, supporting human error detection and enhancing aviation safety.

An explainable longitudinal multi-modal fusion model for predicting neoadjuvant therapy response in women with breast cancer

Multi-modal image analysis using deep learning (DL) lays the foundation for neoadjuvant treatment (NAT) response monitoring. However, existing methods prioritize extracting multi-modal features to enhance predictive performance, with limited consideration on real-world clinical applicability, particularly in longitudinal NAT scenarios with multi-modal data. Here, we propose the Multi-modal Response Prediction (MRP) system, designed to mimic real-world physician assessments of NAT responses in breast cancer. To enhance feasibility, MRP integrates cross-modal knowledge mining and temporal information embedding strategy to handle missing modalities and remain less affected by different NAT settings. We validated MRP through multi-center studies and multinational reader studies. MRP exhibited comparable robustness to breast radiologists, outperforming humans in predicting pathological complete response in the Pre-NAT phase (ΔAUROC 14% and 10% on in-house and external datasets, respectively). Furthermore, we assessed MRP’s clinical utility impact on treatment decision-making. MRP may have profound implications for enrolment into NAT trials and determining surgery extensiveness.

Non-Destructive Pest Detection: Innovations and Challenges in Sensing Airborne Semiochemicals.

Pests, especially invasive ones, pose significant threats to the global ecosystem, crop security, and agriculture economy. Sensing airborne semiochemicals as a nondestructive detection method has been recognized as a promising strategy to detect the presence of these living pests on site. However, sensing airborne semiochemicals in fields is challenging, as they are transmitted in concentrations as low as several nanograms per cubic meter in chemically diverse environments. This low vapor pressure together with similarity in functional groups of pheromones among different species have curtailed the practical deployment of corresponding sensors for real world applications. This review describes the advances in semiochemical detection methods and technologies including traditional analytical instruments, trained animals, and electroantennography with a focus on electronic noses (e-noses). Several key types of volatile organic compound (VOC) sensors used in e-noses are summarized, including their transduction methods, sensing materials, and sensing performance for semiochemical and simulants detection. Notably, it was found that many commercial VOC sensors failed to respond to airborne semiochemicals effectively, leading to a reduced efficiency of e-noses. Future work may focus on developing stable and robust sensing materials with higher sensitivity and selectivity to pheromones and understanding the feasibility of the deployment of the sensors under field conditions.

Enhancing urban infrastructure: water supply and drainage pipeline defect detection with the GLIP pre-trained model and the YOLOv5 model

ABSTRACT Urban water supply and drainage systems are a crucial component of urban infrastructure, directly affecting residents' livelihoods and industrial production. The normal operation of the water supply and the drainage pipeline is of great significance for conserving water resources and preventing water pollution. However, due to characteristics such as deep burial, diverse materials, and extensive lengths, the detection of defects becomes exceptionally complex. Traditional detection methods used in practical applications, such as ground excavation and destructive testing, typically require the shutdown of water pipelines. This process is time-consuming and labor-intensive, often resulting in significant economic losses. This paper proposes an effective technique for detecting defects in the water supply and the drainage pipeline. The method involves capturing images of the inner walls of water supply conduits and subsequently utilizing an artificial intelligence large-scale model approach (grounded language-image pre-training, GLIP) and a You Only Look Once version 5 (YOLOv5) model to detect defects within them. The experimental results show that GLIP demonstrates impressive detection performance in zero-shot scenarios, while YOLOv5 performs well on existing datasets. By combining these two models, we were able to achieve a balance between fast, flexible detection and high precision, making our approach both practical and efficient for real-world applications.