AI Methods Research Articles

Explainable Software Defects Classification Using SMOTE and Machine Learning

Software defect prediction is a critical task in software engineering that aims to identify and mitigate potential defects in software systems. In recent years, numerous techniques and approaches have been developed to improve the accuracy and efficiency of the defect prediction model. In this research paper, we proposed a comprehensive approach that addresses class imbalance by utilizing stratified splitting, explainable AI techniques, and a hybrid machine learning algorithm. To mitigate the impact of class imbalance, we employed stratified splitting during the training and evaluation phases. This method ensures that the class distribution is maintained in both the training and testing sets, enabling the model to learn from and generalize to the minority class examples effectively. Furthermore, we leveraged explainable AI methods, Lime and Shap, to enhance interpretability in the machine learning models. To improve prediction accuracy, we propose a hybrid machine learning algorithm that combines the strength of multiple models. This hybridization allows us to exploit the strength of each model, resulting in improved overall performance. The experiment is evaluated using the NASA-MD datasets. The result revealed that handling the class imbalanced data using stratify splitting approach achieves a better overall performance than the SMOTE approach in Software Defect Detection (SDD).

Neuro – symbolic AI for materials modelling and processes design

This paper presents possibilities of combined use of Artificial Neural Networks (ANNs) with symbolic AI methods for improving the creation and the efficiency of models for materials and manufacturing processes. Besides a short presentation of one of the most basic and common ANNs architecture, which is the Multi-Layer Perceptron (MLP), and the presentation of main symbolic AI methods for knowledge representation and processing, the paper discusses their combined use and possibilities of hybrid neuro–symbolic AI applications for modelling materials and processes. This combined AI modelling approach is planned to be implemented and tested within the Horizon Europe project M2DESCO, which concerns modelling and eco-design of High Entropy Alloys’ (HEAs) coatings.

Deploying AI Methods for Mental Health in Singapore: From Mental Wellness to Serious Mental Health Conditions

Deploying AI Methods for Mental Health in Singapore: From Mental Wellness to Serious Mental Health Conditions

Successful transposition of sperm freezing, in vitro fertilization and artificial insemination methodologies in the Tadjik markhor species (Capra falconeri heptneri).

The objective of the present study was to transpose sperm freezing methodology from domestic goat to the Tadjik markhor (Capra falconeri heptneri) and to address the feasibility to develop IVP and artificial insemination using such frozen semen. Semen of different adult markhor males were successfully recovered by electro-ejaculation and were then frozen using caprine methodology. Frozen semen showed good survival rates at thawing and good fertility rates were assessed in heterologous in vitro fertilization system with goat oocytes. LOPU/IVF was applied for Tadjik markhor females allowing the first successful blastocyst production in vitro. In an applied program, we also transposed successfully intrauterine AI method with frozen/thawed semen to the Tadjik markhor.

Generative AI in Management – Today and Tomorrow

Theoretical background: The rapid and exponential technological advancements have far-reaching impacts on management information systems, management practices, and human life. The promising outcomes in Artificial Intelligence and cutting-edge research on semantic networks and natural language processing have motivated the authors to envision the future of management technology. Purpose of the article: Our paper focuses on the new communication facilities and artificial intelligence models used to process management-type queries in natural language. Research methods: The article discusses recently developed technologies, proposed by Google and Microsoft, notably Google Bard and Bing integrated with ChatGPT-4. Both chatbots use Generative AI methods and large language models to understand domain-based queries and generate answers. Main findings: The practical and social implications of new models in management practice are discussed. To illustrate the qualities and weaknesses of the features of new technologies, four examples of management decision-making are discussed. The case studies also show differences between these two technologies. Finally, the paper concludes by summarizing the expectations and limitations of Generative AI applications in management. The paper is one of the first publications describing and demonstrating the idea of interfaces in natural language in business-oriented applications.

Toward enhancing web privacy on HTTPS traffic: A novel SuperLearner attack model and an efficient defense approach with adversarial examples

Web privacy-enhancing technologies have been grown to preserve the users' privacy against sophisticated traffic analysis attacks aimed at inferring sensitive information. An especially concerning form of traffic analysis attacks is website fingerprinting attacks, allowing local eavesdroppers to identify a user's visited websites over encrypted connections. Leveraging recent advances in AI methods and aligned with the principle of adversarial thinking (think like an adversary), our objective is to better understand privacy vulnerabilities of HTTPS traffic against ever-evolving traffic analysis attacks. Therefore, as a novel contribution, we propose an HTTPS website fingerprinting attack model called Super-Learner Attack, an ensemble of base learners to exploit the strengths and diminish the weaknesses of the individual base learners, including LogisticRegression, DecisionTreeClassifier, Gaussian Naive Bayes, KNeighborsClassifier, AdaboostClassifier, BaggingClassifier, RandomForestClassifier, and ExtraTreeClassifier. The SuperLearner aims to learn fusion weights in a data-adaptive manner to obtain the optimal combination of the base learners. The proposed attack attains a high accuracy of over 97% on HTTPS traffic, outperforming existing attack models. Furthermore, responding to the challenge of website fingerprinting attacks and in direct response to the SuperLearner Attacker, we propose HTTPS Obfuscation Defender. This is a novel and highly effective defense strategy rooted in deception. This strategy disrupts classification by skillfully introducing fake packets into real flows, obfuscating patterns, and disrupting classification. Unlike previous methods, our approach leverages adversarial example algorithms originally designed for image analysis, to generate maximal obfuscation in encrypted HTTPS traffic. It generates perturbations for encrypted HTTPS traffic through the utilization of three well-established adversarial example algorithms. Our experimental results demonstrate that HTTPS Obfuscation Defender significantly reduces the accuracy of website fingerprinting from 97.2% to 2.89%, even when an attacker attempts to adapt to the defense and retrain a classifier using defended traffic. Moreover, our proposed defense ensures a minimal impact on time and bandwidth, making it a highly practical and resource-efficient defense approach.

Converting a DJI Tello Quadcopter into a Face-follower Machine Using the Haar Cascade with PID Controller

Drones have been frequently used for photography in recent years at significantly cheaper rates. However, the most modern drones are exceedingly error-prone and require precise manual control to take high-quality photos or films. We suggest using the AI method of Haar cascades with a PID controller to give drones vision, enabling them to do autonomous tracking and detection. This project aims to improve photography fields. The proposed system tries to detect the face and track the person's movements. This system will help photographers and journalists upgrade their work, even if it is used in surveillance and the military. The algorithm's results show that the DJI Tello tiny drone's camera is capable of detecting and tracking faces. The micro drone was picked since it is lightweight and compact, making its use safe and enabling testing to take place inside. Additionally, the DJI Tello may be easily programmed using Python. The position of the drone is contrasted with the set point in the center of the image to identify errors, allowing control signals for calculating forward/backwards, right/left, and yaw movements. The proposed system results show that the drone can detect and track the face very well, and the PID values are stable.

A Flexibility Platform for Managing Outages and Ensuring the Power System’s Resilience during Extreme Weather Conditions

It is challenging for the European power system to exactly predict RES output and match energy production with demand due to changes in wind and sun intensity and the unavoidable disruptions caused by severe weather conditions. Therefore, in order to address the so-called “flexibility challenge” and implement the variable RES production, the European Union needs flexible solutions. In order to accommodate quicker reactions, compared to those performed today, and the adaptive exploitation of flexibility, grid operators must adjust their operational business model, as the electrical grid transitions from a fully centralized to a largely decentralized system. OneNet aspires to complete this crucial step by setting up a new generation of grid services that can fully utilize distributed generation, storage, and demand responses while also guaranteeing fair, open, and transparent conditions for the consumer. Using AI methods and a cloud-computing approach, the current work anticipates that active management of the power system for TSO–DSO coordination will be improved by the web-based client-server application F-channel. In the current work, a user’s experience with the platform for a Business Use Case (BUC) under the scenario of severe weather conditions is presented. The current work aims to increase the reliability of outage and maintenance plans for the system operators (SOs) by granting them a more accurate insight into the conditions under which the system may be forced to operate in the upcoming period and the challenges that it might face based on those conditions. In this way, the methodology applied in this case could, via AI-driven data exchange and analyses, help SOs change the maintenance and outage plans so the potential grave consequences for the system can be avoided. The SOs will have accurate forecasts of the relevant weather parameters at their disposal that will be used in order to achieve the set targets. The main results of the presented work are that it has a major contribution to the optimal allocation of the available resources, ensures the voltage and frequency stability of the system, and provides an early warning for hazardous power system regimes.

Artificial intelligence-driven new drug discovery targeting serine/threonine kinase 33 for cancer treatment

BackgroundArtificial intelligence (AI) is capable of integrating a large amount of related information to predict therapeutic relationships such as disease treatment with known drugs, gene expression, and drug-target binding. AI has gained increasing attention as a promising tool for next-generation drug development.MethodsAn AI method was used for drug repurposing and target identification for cancer. Among 8 survived candidates after background checking, N-(1-propyl-1H-1,3-benzodiazol-2-yl)-3-(pyrrolidine-1-sulfonyl) benzamide (Z29077885) was newly selected as an new anti-cancer drug, and the anti-cancer efficacy of Z29077885 was confirmed using cell viability, western blot, cell cycle, apoptosis assay in MDA-MB 231 and A549 in vitro. Then, anti-tumor efficacy of Z29077885 was validated in an in vivo A549 xenograft in BALB/c nude mice.ResultsFirst, we discovered an antiviral agent, Z29077885, as a new anticancer drug candidate using the AI deep learning method. Next, we demonstrated that Z29077885 inhibits Serine/threonine kinase 33 (STK33) enzymatic function in vitro and showed the anticancer efficacy in various cancer cells. Then, we found enhanced apoptosis via S-phase cell cycle arrest as the mechanism underlying the anticancer efficacy of Z29077885 in both lung and breast cancer cells. Finally, we confirmed the anti-tumor efficacy of Z29077885 in an in vivo A549 xenograft.ConclusionsIn this study, we used an AI-driven screening strategy to find a novel anticancer medication targeting STK33 that triggers cancer cell apoptosis and cell cycle arrest at the s phase. It will pave a way to efficiently discover new anticancer drugs.

Software Testing Based on Research: A Road Map

This paper serves as a guide for both researchers and students who are new to the research area of Search-Based Package Testing. The application of metaheuristic explore methods in this context specifically pertains to utilizing these algorithms for generating test data. In the realm of software engineering research, software testing emerges as a robust and fertile ground for exploration. The integration of AI methods into software program testing is an evolving research direction. Often, newcomers to this field face challenges due to limited knowledge about the interaction between software testing and artificial intelligence. This paper aims to provide a roadmap for these new researchers or students in the field.

Advancing Bridge Structural Health Monitoring: Insights into Knowledge-Driven and Data-Driven Approaches

Structural health monitoring (SHM) is increasingly being used in the field of bridge engineering, and the technology for monitoring bridges has undergone a radical change. It has evolved from the initial local monitoring and assessment, which relied mainly on manual work, to the current all-round and full-time intelligent assessment provided by intelligent monitoring systems. This paper reviews the development of SHM technology in the civil engineering field and examines two current artificial intelligence (AI) methods in bridge SHM, namely knowledge-driven and data-driven approaches. The advantages and disadvantages of these two AI methods are analyzed, and future development trends are also discussed. The overview results reveal that knowledge-driven methods have the advantages of interpretability and stability. However, their current application is limited, and significant technical bottlenecks remain. On the other hand, the data-driven approach demonstrates higher efficiency and accuracy. Nevertheless, it is characterized by instability and insecurity due to its "black-box" nature, which hinders its ability to explain the internal operation mechanism. Given these findings, the hybrid knowledge-data-driven approach emerges as a potential solution. This approach can effectively integrate the advantages of both knowledge-driven and data-driven methods while avoiding their respective disadvantages. Consequently, the hybrid approach proves to be more stable, safe, and efficient in practical applications.

Artificial Intelligence and Machine Learning (AIML) approaches to empower genomics, drug, and biomarker discovery in ADRD

AbstractBackgroundIn 2021, the NIA launched the Artificial Intelligence and Machine Learning (AIML) Consortium, to develop novel AI and machine learning techniques to tackle pressing questions in Alzheimer’s disease research.MethodDrawing on vast repositories of data from the Alzheimer’s Disease Sequencing Project (ADSP) and other large public biobanks, the 8 awardees funded under this program meet monthly to present their innovations in data analytics in 4 main areas: (1) genome‐guided drug discovery and drug repositioning for AD; (2) discovery of novel genetic and molecular targets for AD therapeutics using whole genome and omics data; and (3) understanding the regulatory code of the genome using deep learning and bioinformatics, and (4) use of electronic health records (EHR) to study treatment outcomes in ADRD. A second branch of AIML activity draws on repositories of neuroimaging data (MRI/FLAIR, DTI, and amyloid‐ and tau‐sensitive PET) to (1) identify ADRD subtypes based on neuroimaging, clinical and neuropsychological data, that differ in terms of their genomic drivers and expected treatment response (for precision medicine and drug trial stratification), (2) perform genetic analysis of endophenotypes derived from brain images (such as longitudinal hippocampal volume changes, atrophy subtype scores, and PET or WMH signatures) for biomarker‐guided gene discovery, and (3) predict neuropathology (TDP‐43, CAA, LBD) from multimodal imaging using AI. The AIML initiative benefits from and helps the ADSP’s Phenotype Harmonization Consortium, as AI methods can also be used to adjust neuroimaging data for site and protocol effects (harmonization).ResultIn this overview of the AIML Consortium, we highlight some recent successes using (1) convolutional neural networks (CNNs), GANs and deep learning for ADRD subtyping and staging; (2) discovery of AD‐relevant motifs and networks in whole genome sequences and functional genomics data; (3) prognostic models of clinical outcomes based on WGS and biomarkers; and (4) novel methods from statistical physics, based on evolutionary action, which model the fitness effect of coding mutations analytically.ConclusionAIML investigators are hosting regular workshops at AAIC and upcoming genomics and clinical neurology conferences to disseminate and teach these approaches to the AD and broader scientific community.

Theoretical Analyses of Multiobjective Evolutionary Algorithms on Multimodal Objectives.

Multiobjective evolutionary algorithms are successfully applied in many real-world multiobjective optimization problems. As for many other AI methods, the theoretical understanding of these algorithms is lagging far behind their success in practice. In particular, previous theory work considers mostly easy problems that are composed of unimodal objectives. As a first step towards a deeper understanding of how evolutionary algorithms solve multimodal multiobjective problems, we propose the OneJumpZeroJump problem, a bi-objective problem composed of two objectives isomorphic to the classic jump function benchmark. We prove that the simple evolutionary multiobjective optimizer (SEMO) with probability one does not compute the full Pareto front, regardless of the runtime. In contrast, for all problem sizes n and all jump sizes k∈[4..n2-1], the global SEMO (GSEMO) covers the Pareto front in an expected number of Θ((n-2k)nk) iterations. For k=o(n), we also show the tighter bound 32enk+1±o(nk+1), which might be the first runtime bound for an MOEA that is tight apart from lower-order terms. We also combine the GSEMO with two approaches that showed advantages in single-objective multimodal problems. When using the GSEMO with a heavy-tailed mutation operator, the expected runtime improves by a factor of at least kΩ(k). When adapting the recent stagnation-detection strategy of Rajabi and Witt (2022) to the GSEMO, the expected runtime also improves by a factor of at least kΩ(k) and surpasses the heavy-tailed GSEMO by a small polynomial factor in k. Via an experimental analysis, we show that these asymptotic differences are visible already for small problem sizes: A factor-5 speed-up from heavy-tailed mutation and a factor-10 speed-up from stagnation detection can be observed already for jump size 4 and problem sizes between 10 and 50. Overall, our results show that the ideas recently developed to aid single-objective evolutionary algorithms to cope with local optima can be effectively employed also in multiobjective optimization.

Rapid intraoperative multi-molecular diagnosis of glioma with ultrasound radio frequency signals and deep learning

Molecular diagnosis is crucial for biomarker-assisted glioma resection and management. However, some limitations of current molecular diagnostic techniques prevent their widespread use intraoperatively. With the unique advantages of ultrasound, this study developed a rapid intraoperative molecular diagnostic method based on ultrasound radio-frequency signals. We built a brain tumor ultrasound bank with 169 cases enrolled since July 2020, of which 43483 RF signal patches from 67 cases with a pathological diagnosis of glioma were a retrospective cohort for model training and validation. IDH1 and TERT promoter (TERTp) mutations and 1p/19q co-deletion were detected by next-generation sequencing. We designed a spatial-temporal integration model (STIM) to diagnose the three molecular biomarkers, thus establishing a rapid intraoperative molecular diagnostic system for glioma, and further analysed its consistency with the fifth edition of the WHO Classification of Tumors of the Central Nervous System (WHO CNS5). We tested STIM in 16-case prospective cohorts, which contained a total of 10384 RF signal patches. Two other RF-based classical models were used for comparison. Further, we included 20 cases additional prospective data for robustness test (ClinicalTrials.govNCT05656053). In the retrospective cohort, STIM achieved a mean accuracy and AUC of 0.9190 and 0.9650 (95% CI, 0.94-0.99) respectively for the three molecular biomarkers, with a total time of 3s and a 96% match to WHO CNS5. In the prospective cohort, the diagnostic accuracy of STIM is 0.85±0.04 (mean±SD) for IDH1, 0.84±0.05 for TERTp, and 0.88±0.04 for 1p/19q. The AUC is 0.89±0.02 (95% CI, 0.84-0.94) for IDH1, 0.80±0.04 (95% CI, 0.71-0.89) for TERTp, and 0.85±0.06 (95% CI, 0.73-0.98) for 1p/19q. Compared to the second best available method based on RF signal, the diagnostic accuracy of STIM is improved by 16.70% and the AUC is improved by 19.23% on average. STIM is a rapid, cost-effective, and easy-to-manipulate AI method to perform real-time intraoperative molecular diagnosis. In the future, it may help neurosurgeons designate personalized surgical plans and predict survival outcomes. A full list of funding bodies that contributed to this study can be found in the Acknowledgements section.

Enabling the clinical application of artificial intelligence in genomics: a perspective of the AMIA Genomics and Translational Bioinformatics Workgroup.

Given the importance AI in genomics and its potential impact on human health, the American Medical Informatics Association-Genomics and Translational Biomedical Informatics (GenTBI) Workgroup developed this assessment of factors that can further enable the clinical application of AI in this space. A list of relevant factors was developed through GenTBI workgroup discussions in multiple in-person and online meetings, along with review of pertinent publications. This list was then summarized and reviewed to achieve consensus among the group members. Substantial informatics research and development are needed to fully realize the clinical potential of such technologies. The development of larger datasets is crucial to emulating the success AI is achieving in other domains. It is important that AI methods do not exacerbate existing socio-economic, racial, and ethnic disparities. Genomic data standards are critical to effectively scale such technologies across institutions. With so much uncertainty, complexity and novelty in genomics and medicine, and with an evolving regulatory environment, the current focus should be on using these technologies in an interface with clinicians that emphasizes the value each brings to clinical decision-making.

Intelligent transportation systems as cyberphysical systems in transport

In paper we offer a report on the results of the comparative analysis of existing concepts of the functioning and development of Smart Logistic. In paper shown that at present ITS is the most developed concept for the implementation of Smart Logistic and, in its essence, approaches the essence of the CPS concept as intelligent automatic (or close to automatic) control systems for physical objects and processes of various nature. Within the framework of the CPS concept, through test studies using various AI methods (GA, ESA, AA and ACAmod), it was determined that the most effective method for optimizing the route of freight traffic is ACAmod, which allows dynamic routing of logistics flows in real time, taking into account the non-stationary dynamics of TF. One of the promising areas for further research, the authors consider the creation of an intelligent system for supporting decisions on transport and logistics management of freight traffic in real time taking into account the influence of external factors of various nature on the transportation processes. The creation of such a system should be based on a modern interpretation of the concept of ITS as a CPS for Smart Logistics. In addition, in this report within the concept of the Cyber-Physical Systems are show the results of determining of the most effective method for optimizing of logistics flows in real time, taking into account the non-stationary dynamics of traffic flows

Evaluating Transparency and Explainability in AI-Driven Planning and Scheduling: A Comprehensive Literature Review

As AI technologies rapidly integrate into our work environments, the need for systems that effectively collaborate with humans grows. A crucial aspect of this interaction is the explainability of AI behavior to human users. While AI has surpassed human decision-making abilities in many areas, its opaque algorithms often produce accurate yet incomprehensible results. This lack of transparency can hinder trust between humans and machines. The TUPLES project, a three-year Horizon Europe research initiative, seeks to address this by developing AI-based planning and scheduling tools with a human-centered focus. By utilizing both data-driven and symbolic AI methods, TUPLES aims to create scalable, transparent, robust, and secure systems. The project employs a use-case-oriented approach, drawing from industry expertise to ensure relevance in areas such as manufacturing, aviation, and waste management. The EU guidelines for Trustworthy AI emphasize key principles like human oversight, transparency, fairness, societal well-being, and accountability. The Assessment List for Trustworthy Artificial Intelligence (ALTAI) serves as a selfassessment tool for evaluating AI systems. Current planning and scheduling tools only partially meet these criteria, highlighting the need for innovative AI development. Our literature review explored the transparency and explainability of AI algorithms in planning and scheduling, identifying metrics and recommendations. The study underscores the role of Explainable AI (XAI) in addressing the black box problem, using techniques that make AI systems more understandable and accurate. XAI relies on human explanation concepts to ensure a human-centered approach, offering specific methods and guidance for selecting appropriate tools in planning and scheduling contexts.

Explainable artificial intelligence for omics data: a systematic mapping study.

Researchers increasingly turn to explainable artificial intelligence (XAI) to analyze omics data and gain insights into the underlying biological processes. Yet, given the interdisciplinary nature of the field, many findings have only been shared in their respective research community. An overview of XAI for omics data is needed to highlight promising approaches and help detect common issues. Toward this end, we conducted a systematic mapping study. To identify relevant literature, we queried Scopus, PubMed, Web of Science, BioRxiv, MedRxiv and arXiv. Based on keywording, we developed a coding scheme with 10 facets regarding the studies' AI methods, explainability methods and omics data. Our mapping study resulted in 405 included papers published between 2010 and 2023. The inspected papers analyze DNA-based (mostly genomic), transcriptomic, proteomic or metabolomic data by means of neural networks, tree-based methods, statistical methods and further AI methods. The preferred post-hoc explainability methods are feature relevance (n= 166) and visual explanation (n= 52), while papers using interpretable approaches often resort to the use of transparent models (n= 83) or architecture modifications (n= 72). With many research gaps still apparent for XAI for omics data, we deduced eight research directions and discuss their potential for the field. We also provide exemplary research questions for each direction. Many problems with the adoption of XAI for omics data in clinical practice are yet to be resolved. This systematic mapping study outlines extant research on the topic and provides research directions for researchers and practitioners.

What Is Hidden in Clear Sight and How to Find It—A Survey of the Integration of Artificial Intelligence and Eye Tracking

This paper presents an overview of the uses of the combination of eye tracking and artificial intelligence. In the paper, several aspects of both eye tracking and applied AI methods have been analyzed. It analyzes the eye tracking hardware used along with the sampling frequency, the number of test participants, additional parameters, the extraction of features, the artificial intelligence methods used and the methods of verification of the results. Finally, it includes a comparison of the results obtained in the analyzed literature and a discussion about them.

Analysis of Estimation of Soundness and Deterioration Factors of Sewage Pipes Using Machine Learning

In Japan, there are a massive number of sewage pipes buried in the ground. In order to operate sustainable sewerage systems, it is necessary to estimate the soundness of sewage pipes accurately and to conduct repairs and other measures according to the soundness of the pipes. In previous studies, statistical and machine learning methods have been used to estimate the soundness of sewage pipes, but all of these studies formulated the soundness of sewage pipes as a binary classification problem (e.g., good or poor). In contrast, this study attempted to predict the soundness of sewage pipes in more detail by setting up four classes of pipe soundness. Inspection data of sewage pipes in City A were used as training data, and XGBoost was used as the machine learning model. Machine learning models have a high prediction performance, but the uncertainty of the prediction basis is an issue. In this study, SHAP (Shapley additive explanations), an Explainable AI method, was used to interpret the model to clarify the influence of sewer pipe specifications (e.g., pipe age) and topographical specifications (e.g., annual precipitation) on the prediction, and to extract deterioration factors. By interpreting the model using SHAP, it was possible to quantify whether factors such as pipe age and pipe length have a positive or negative impact on the deterioration of sewage pipes. Previous studies using machine learning methods have not clarified whether factors have a positive or negative effect on deterioration. The knowledge on deterioration factors obtained in this study may provide useful information for the sustainable operation of sewage systems.