Intelligence-based Approach Research Articles

Abstract TP162: Artificial Intelligence-Based Approach for Localization of Intracranial Aneurysms on Head MRI Images

Background: Detection of intracranial aneurysms (IAs) is a time consuming and error prone process. Therefore, solutions that can localize IAs with a high sensitivity are required. Several artificial intelligence (AI)-based automated diagnoses on medical images have recently been reported. We aimed to develop an automated diagnosis system for the location and maximum diameter of IAs on MRI images in the present study. Methods: In 1310 patients with or without IAs, 937 MRIs were used for training data and 373 cases for test data. The definition of the correct diagnosis of the location was that the center of the aneurysm diagnosed by the AI system is within the area of the IA of ground truth. The nnU-Net was used for the deep learning. Developed AI system was verified with 5-fold cross validation with test data. The maximum diameter was automatically calculated from extracted domes of IAs. Results: Of the 937 patients used for the model development, 778 (83%) had IAs including 146 patients (19%) with multiple aneurysms and 159 (17%) had no IA. In total 1213 IAs, 78% were small aneurysms of 2 to 5 mm. Of the 373 cases for the validation, 17 (4.5%) had IAs, which is close to the real-world setting. Internal validation of the developed diagnosis model for IA locations showed high efficiency of AUC = 0.92, sensitivity = 15/17 (88%), and false negative rate = 0.44 aneurysms/person. The diameter diagnosis model for the maximum diameter also achieved high accuracy within 1.2 mm of mean absolute error. Conclusion: We successfully developed a powerful diagnosis model for IAs with AI technique. The present model potentially reduces human effort required for the IA diagnosis.

Artificial intelligence-based approach for water-cut prediction in crude oil processing

Artificial intelligence-based approach for water-cut prediction in crude oil processing

An artificial intelligence-based approach for avoiding traffic congestion in connected autonomous vehicles

An artificial intelligence-based approach for avoiding traffic congestion in connected autonomous vehicles

ImpACT Project: Improving Access to Clinical Trials in Victoria, an Artificial Intelligence-Based Approach.

Enhancing the speed and efficiency of clinical trial recruitment is a key objective across international health systems. This study aimed to use artificial intelligence (AI) applied in the Victorian Cancer Registry (VCR), a population-based cancer registry, to assess (1) if VCR received all relevant pathology reports for three clinical trials, (2) AI accuracy in auto-extracting information from pathology reports for recruitment, and (3) the number of participants approached for trial enrollment using the AI approach compared with standard hospital-based recruitment. To verify pathology report accessibility for VCR trial enrollment, reports from the laboratory were cross-referenced. To determine the accuracy of a Rapid Case Ascertainment (RCA) module of the AI software in extracting key clinical variables from the pathology report, data were compared with manually reviewed reports. To examine the effectiveness of the AI recruitment approach, the number of patients approached for recruitment was compared with standard practice. Of the 195 reports provided by the pathology laboratory, 185 (94.9%) were received by VCR, 73 of 195 (37.4%) were eligible for the studies, and 5 of 73 (6.8%) eligible cases had not been received by the VCR. The RCA module demonstrated an accuracy of 93% and an F1 score of 0.94 in extracting key clinical variables. However, the RCA false-positive rate was 10% and the false-negative rate was 5%. The standard hospital approach selected fewer cases for approach to clinical trials compared with the RCA module approach, 8 of 336 (2.4%) versus 12 of 336 (3.6%), respectively. Using AI to screen potentially eligible cases for recruitment to three clinical trials resulted in a 50% increase in eligible cases being approached for enrollment.

Machine learning-based pipeline for automated intracerebral hemorrhage and drain detection, quantification, and classification in non-enhanced CT images (NeuroDrAIn).

External drainage represents a well-established treatment option for acute intracerebral hemorrhage. The current standard of practice includes post-operative computer tomography imaging, which is subjectively evaluated. The implementation of an objective, automated evaluation of postoperative studies may enhance diagnostic accuracy and facilitate the scaling of research projects. The objective is to develop and validate a fully automated pipeline for intracerebral hemorrhage and drain detection, quantification of intracerebral hemorrhage coverage, and detection of malpositioned drains. In this retrospective study, we selected patients (n = 68) suffering from supratentorial intracerebral hemorrhage treated by minimally invasive surgery, from years 2010-2018. These were divided into training (n = 21), validation (n = 3) and testing (n = 44) datasets. Mean age (SD) was 70 (±13.56) years, 32 female. Intracerebral hemorrhage and drains were automatically segmented using a previously published artificial intelligence-based approach. From this, we calculated coverage profiles of the correctly detected drains to quantify the drains' coverage by the intracerebral hemorrhage and classify malpositioning. We used accuracy measures to assess detection and classification results and intraclass correlation coefficient to assess the quantification of the drain coverage by the intracerebral hemorrhage. In the test dataset, the pipeline showed a drain detection accuracy of 0.97 (95% CI: 0.92 to 0.99), an agreement between predicted and ground truth coverage profiles of 0.86 (95% CI: 0.85 to 0.87) and a drain position classification accuracy of 0.88 (95% CI: 0.77 to 0.95) resulting in area under the receiver operating characteristic curve of 0.92 (95% CI: 0.85 to 0.99). We developed and statistically validated an automated pipeline for evaluating computed tomography scans after minimally invasive surgery for intracerebral hemorrhage. The algorithm reliably detects drains, quantifies drain coverage by the hemorrhage, and uses machine learning to detect malpositioned drains. This pipeline has the potential to impact the daily clinical workload, as well as to facilitate the scaling of data collection for future research into intracerebral hemorrhage and other diseases.

Change Meat Resistance: Systematic Literature Review on Consumer Resistance to the Alternative Protein Transition.

Consumer resistance significantly impedes the transition from animal-derived proteins to alternative protein sources through a dual mechanism: consumer reluctance to change entrenched meat-eating habits and hesitation among policymakers, marketers, and practitioners due to anticipated resistance. The concept of resistance is intricate and viewed diversely across research disciplines. We conducted an extensive systematic literature review supplemented by an artificial intelligence-based approach. We evaluated 3,387 studies to identify 51 key papers. The results reveal that resistance is under-researched. Resistance to plant-based diets is associated with practical barriers, whereas resistance to reducing meat consumption is tied more to moral and social aspects. Resistance predominantly manifests among unmotivated meat lovers with strong meat-eating habits. On a positive note, resistance reflects consumer concerns, tends to diminish over time, is overrated, and is specifically linked to specific consumer groups. Thus, addressing resistance is vital, as it facilitates the transition to a more sustainable and healthy food supply that relies less on animal proteins.

A deep learning framework deploying segment anything to detect pan-cancer mitotic figures from haematoxylin and eosin-stained slides

Mitotic activity is an important feature for grading several cancer types. However, counting mitotic figures (cells in division) is a time-consuming and laborious task prone to inter-observer variation. Inaccurate recognition of MFs can lead to incorrect grading and hence potential suboptimal treatment. This study presents an artificial intelligence-based approach to detect mitotic figures in digitised whole-slide images stained with haematoxylin and eosin. Advances in this area are hampered by the small size and variety of datasets available. To address this, we create the largest dataset of mitotic figures (N = 74,620), combining an in-house dataset of soft tissue tumours with five open-source datasets. We then employ a two-stage framework, named the Optimised Mitoses Generator Network (OMG-Net), to identify mitotic figures. This framework first deploys the Segment Anything Model to automatically outline cells, followed by an adapted ResNet18 that distinguishes mitotic figures. OMG-Net achieves an F1 score of 0.84 in detecting pan-cancer mitotic figures, including human breast carcinoma, neuroendocrine tumours, and melanoma. It outperforms previous state-of-the-art models in hold-out test sets. To summarise, our study introduces a generalisable data creation and curation pipeline and a high-performance detection model, which can largely contribute to the field of computer-aided mitotic figure detection.

Novel artificial intelligence-based identification of drug-gene-disease interaction using protein-protein interaction

The evaluation of drug-gene-disease interactions is key for the identification of drugs effective against disease. However, at present, drugs that are effective against genes that are critical for disease are difficult to identify. Following a disease-centric approach, there is a need to identify genes critical to disease function and find drugs that are effective against them. By contrast, following a drug-centric approach comprises identifying the genes targeted by drugs, and then the diseases in which the identified genes are critical. Both of these processes are complex. Using a gene-centric approach, whereby we identify genes that are effective against the disease and can be targeted by drugs, is much easier. However, how such sets of genes can be identified without specifying either the target diseases or drugs is not known. In this study, a novel artificial intelligence-based approach that employs unsupervised methods and identifies genes without specifying neither diseases nor drugs is presented. To evaluate its feasibility, we applied tensor decomposition (TD)-based unsupervised feature extraction (FE) to perform drug repositioning from protein-protein interactions (PPI) without any other information. Proteins selected by TD-based unsupervised FE include many genes related to cancers, as well as drugs that target the selected proteins. Thus, we were able to identify cancer drugs using only PPI. Because the selected proteins had more interactions, we replaced the selected proteins with hub proteins and found that hub proteins themselves could be used for drug repositioning. In contrast to hub proteins, which can only identify cancer drugs, TD-based unsupervised FE enables the identification of drugs for other diseases. In addition, TD-based unsupervised FE can be used to identify drugs that are effective in in vivo experiments, which is difficult when hub proteins are used. In conclusion, TD-based unsupervised FE is a useful tool for drug repositioning using only PPI without other information.

Artificial Intelligence-Based Approach for Damage Localization in Ultrasonic Guided Wave-Based Structural Health Monitoring

The continuous monitoring of structural integrity is crucial, as imperceptible damage may appear at any point throughout a structure's lifespan. Several Structural Health Monitoring (SHM) technologies have been developed so far to detect and assess defects in structures. Deep Learning is a common tool for processing data obtained with SHM systems. Although many artificial intelligence-based SHM technologies exist already for damage detection, only a few are focused on damage localization. Existing localization approaches are limited through them being applied on defined simple structures and requiring a large amount of data, which is usually unavailable in practical applications. In this work, measured data from guided ultrasonic wave propagation is used to determine the location of damage in a composite stiffened structure representative of fuselage segments. A novel artificial intelligencebased approach for damage localization is presented and tested with a feed-forward as well as a convolutional neural network. Both architectures show that localization is possible. The accuracy is then analyzed with the probability of localization method and compared to existing non-artificial intelligence-based approaches. These results make it possible to define the minimum damage that can be correctly localized.

BCG therapy for bladder cancer: Exploring patient experiences and concerns through artificial intelligence-based social media analysis

Background There is a notable disparity between the guidelines for BCG therapy in non-muscle invasive bladder cancer (NMIBC). Reddit has emerged as a popular online platform for individuals seeking information and exchanging their experiences related to bladder cancer. Objective To investigate and classify public opinions about intravesical BCG therapy as shared on Reddit, a popular social media platform. Methods This study employed an artificial intelligence-based approach to examine discussions related to intravesical BCG therapy on a social media platform over the past ten years. An artificial intelligence framework was developed to categorize these conversations into distinct topics and thematic categories. This framework included a partially supervised model for processing natural language (using BERT [Bidirectional Encoder Representations from Transformers]), a method for reducing data complexity, and an algorithm for clustering. Additionally, each conversation was assessed for sentiment. Results A total of 1223 unique discussions related to BCG therapy were analyzed, comprising 110 unique posts and 1113 comments from 268 distinct authors. We identified four overarching thematic groups: 1) BCG administration procedures, (2) hesitancy in initiating or maintaining BCG treatment, (3) issues related to BCG shortage and alternative treatments, and (4) side effects of BCG treatment. Sentiment analysis of the 1223 discussions revealed that 25.2% (308) exhibited a negative sentiment, 58.3% (713) were neutral, and 16.5% (202) showed a positive sentiment. Conclusion Online social media often contains detailed personal experiences with BCG therapy, not commonly found in medical literature. Understanding these experiences can help medical professionals improve care and treatment adherence in NMIBC.

Enhanced interpretable thyroid disease diagnosis by leveraging synthetic oversampling and machine learning models

Thyroid illness encompasses a range of disorders affecting the thyroid gland, leading to either hyperthyroidism or hypothyroidism, which can significantly impact metabolism and overall health. Hypothyroidism can cause a slowdown in bodily processes, leading to symptoms such as fatigue, weight gain, depression, and cold sensitivity. Hyperthyroidism can lead to increased metabolism, causing symptoms like rapid weight loss, anxiety, irritability, and heart palpitations. Prompt diagnosis and appropriate treatment are crucial in managing thyroid disorders and improving patients’ quality of life. Thyroid illness affects millions worldwide and can significantly impact their quality of life if left untreated. This research aims to propose an effective artificial intelligence-based approach for the early diagnosis of thyroid illness. An open-access thyroid disease dataset based on 3,772 male and female patient observations is used for this research experiment. This study uses the nominal continuous synthetic minority oversampling technique (SMOTE-NC) for data balancing and a fine-tuned light gradient booster machine (LGBM) technique to diagnose thyroid illness and handle class imbalance problems. The proposed SNL (SMOTE-NC-LGBM) approach outperformed the state-of-the-art approach with high-accuracy performance scores of 0.96. We have also applied advanced machine learning and deep learning methods for comparison to evaluate performance. Hyperparameter optimizations are also conducted to enhance thyroid diagnosis performance. In addition, we have applied the explainable Artificial Intelligence (XAI) mechanism based on Shapley Additive exPlanations (SHAP) to enhance the transparency and interpretability of the proposed method by analyzing the decision-making processes. The proposed research revolutionizes the diagnosis of thyroid disorders efficiently and helps specialties overcome thyroid disorders early.

Intelligent-based approach for parameters identification and control of fractional order systems

The optimization algorithms for identification and control processes take more and more place in the industrial domain. However, determining a control law for fractional systems remains challenging for researchers today. This paper presents a simple method for using a fractional PID (FOPID) corrector to regulate a class of fractional systems that show a stable step response. The procedure is based on an algorithmic identification to assimilate the fractional system into a simple model. Optimization methods were used. Using techniques like Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Artificial Bee Colony (ABC), and Ant Colony Optimization (ACO) to ascertain the parameters of the basic model, a first-order system with delay. Then, the FOPID controller will be calculated using Ziegler–Nichols methods and optimization algorithms to stabilize the real process. Better performance Overshoot and Settling time as well as minimal Energy control effort are benefits of the proposed design. To confirm that the fractional regulator is robust which is optimized by the best-chosen method (PSO), the parameters of the chosen process will be randomly varied. Through two numerical simulations, the efficiency of the suggested method has been confirmed.

Ensemble-Based Alzheimer's Disease Classification Using Features Extracted from Hog Descriptor and Pre-trained Models

Alzheimer's Disease is the most common type of dementia and is a progressive, neurodegenerative disease. The disease worsens over time, and the patient becomes bedridden, unable to move or understand what is happening around him. The main concern of medicine is to slow down the progression of the disease for which no treatment has yet been developed. Artificial intelligence studies have achieved significant success in detecting many diseases. In this study, an artificial intelligence-based approach that uses MR images of the early stage of Alzheimer's Disease to detect the disease at an early stage is presented. Initially, a new dataset was created through the application of the fuzzy technique, thereby expanding the feature space. Then, an ensemble learning-based hybrid deep learning model was developed to reduce the misclassification rate for all classes. The features derived from the inception module, residual modules, and histogram of oriented gradients descriptor are subjected to classification through bagging and boosting algorithms. The proposed model has surpassed many state-of-the-art studies by achieving a high success rate of 99.60% in detecting Alzheimer's disease in its early stages.

Acute cholecystitis diagnosis in the emergency department: an artificial intelligence-based approach.

This study aimed to assess the diagnostic performance of a support vector machine (SVM) algorithm for acute cholecystitis and evaluate its effectiveness in accurately diagnosing this condition. Using a retrospective analysis of patient data from a single center, individuals with abdominal pain lasting one week or less were included. The SVM model was trained and optimized using standard procedures. Model performance was assessed through sensitivity, specificity, accuracy, and AUC-ROC, with probability calibration evaluated using the Brier score. Among 534 patients, 198 (37.07%) were diagnosed with acute cholecystitis. The SVM model showed balanced performance, with a sensitivity of 83.08% (95% CI: 71.73-91.24%), a specificity of 80.21% (95% CI: 70.83-87.64%), and an accuracy of 81.37% (95% CI: 74.48-87.06%). The positive predictive value (PPV) was 73.97% (95% CI: 65.18-81.18%), the negative predictive value (NPV) was 87.50% (95% CI: 80.19-92.37%), and the AUC-ROC was 0.89 (95% CI: 0.85 to 0.93). The Brier score indicated well-calibrated probability estimates. The SVM algorithm demonstrated promising potential for accurately diagnosing acute cholecystitis. Further refinement and validation are needed to enhance its reliability in clinical practice.

Chirped apodized fiber Bragg gratings inverse design via deep learning

Overcoming inverse design problems in fiber Bragg gratings (FBGs) can be challenging due to the significant nonlinearity of the problem and the intricate relationship between structural properties and optical characteristics. Here, we present a novel artificial intelligence-based approach that effectively addresses these challenges. We introduce a methodology centered on applying deep learning (DL) to estimate the reflective spectrum of FBGs. The results highlight DL’s exceptional capability in designing chirped apodized FBGs, with our model demonstrating significantly enhanced computational efficiency relative to traditional numerical simulations. Notably, our DL-based approach exhibits the remarkable ability to tackle the inverse design challenges of FBGs, thereby eliminating the reliance on trial-and-error or empirical methodologies. The predictive losses for both the forward and inverse models are impressively minimal, with low loss values of 2.2 × 10-2 and 1.6 × 10-2, respectively. The FBG configurations derived via DL are ideally suited for optical communications, heralding significant advancements in all-optical signal processing.

Accurate diagnosis of acute appendicitis in the emergency department: an artificial intelligence-based approach.

The diagnosis of abdominal pain in emergency departments is challenging, and appendicitis is a common concern. Atypical symptoms often delay diagnosis. Although the Alvarado score aids in decision-making, its low specificity can lead to unnecessary surgeries. By leveraging machine learning, we aim to enhance diagnostic accuracy by predicting appendicitis and distinguishing it from other causes of abdominal pain in the emergency department. Data were collected from 534 patients who presented with acute abdominal pain. Patient characteristics, laboratory results, and causes of pain were recorded. Machine learning algorithms (support vector classifier, random forest classifier, gradient boosting classifier, and Gaussian naive Bayes) were used to predict the cause of pain. Model calibration was assessed using the Brier score. The mean age was 46.89 (20.3) years, with an almost equal sex distribution (49% male, 51% female). Cholecystitis was the most prevalent outcome (37.07%), followed by appendicitis (25.84%). The Gaussian naive Bayes model exhibited superior performance in terms of accuracy (95.03% 95% CI 90.44-97.83%), sensitivity (87.18% 95% CI 72.57-95.70%), and specificity (97.54% 95% CI 92.98-99.49%), while the random forest model showed a sensitivity of 79.49%, specificity of 96.72%, and accuracy of 92.55%. The gradient boosting algorithm achieved a sensitivity, specificity, and accuracy of 89.74%, 95.90%, and 94.41%, respectively. The support vector classifier demonstrated a sensitivity of 89.74%, specificity of 92.62%, and accuracy of 91.93%. The use of modern machine learning methods aids in the accurate diagnosis of appendicitis.

Artificial intelligence-based approach to detect malicious users using deep learning and optimization techniques

Artificial intelligence-based approach to detect malicious users using deep learning and optimization techniques

An improved deep convolutional neural network-based YouTube video classification using textual features

Video content on the web platform has increased explosively during the past decade, thanks to the open access to Facebook, YouTube, etc. YouTube is the second-largest social media platform nowadays containing more than 37 million YouTube channels. YouTube revealed at a recent press event that 30,000 new content videos per hour and 720,000 per day are posted. There is a need for an advanced deep learning-based approach to categorize the huge database of YouTube videos. This study aims to develop an artificial intelligence-based approach to categorize YouTube videos. This study analyzes the textual information related to videos like titles, descriptions, user tags, etc. using YouTube exploratory data analysis (YEDA) and shows that such information can be potentially used to categorize videos. A deep convolutional neural network (DCNN) is designed to categorize YouTube videos with efficiency and high accuracy. In addition, recurrent neural network (RNN), and gated recurrent unit (GRU) are also employed for performance comparison. Moreover, logistic regression, support vector machines, decision trees, and random forest models are also used. A large dataset with 9 classes is used for experiments. Experimental findings indicate that the proposed DCNN achieves the highest receiver operating characteristics (ROC) area under the curve (AUC) score of 99% in the context of YouTube video categorization and 96% accuracy which is better than existing approaches. The proposed approach can be used to help YouTube users suggest relevant videos and sort them by video category.

Application of an Intelligence-Based Approach in Learning Islamic Religion

Berdasarkan hasil observasi yang dilakukan di SMA N 1 2X11 Enam Lingkung, hasil belajar siswa masih rendah, terbukti dengan masih banyaknya siswa yang belum menyelesaikan pembelajaran agama Islam. Selain itu, aktivitas siswa dalam mengikuti proses pembelajaran masih rendah. Hal ini disebabkan karena kecerdasan intrapersonal dan kecerdasan interpersonal siswa belum berkembang dengan baik dalam pembelajaran agama Islam. Untuk mengatasi permasalahan tersebut, peneliti menerapkan “pendekatan berbasis kecerdasan dalam pembelajaran agama Islam”. Pendekatan berbasis kecerdasan merupakan pendekatan pembelajaran yang bertujuan untuk mengaktifkan kecerdasan siswa melalui metode pembelajaran yang bervariasi. Oleh karena itu, rumusan masalahnya adalah apakah hasil belajar siswa pada pendekatan berbasis kecerdasan lebih tinggi dibandingkan dengan hasil belajar siswa pada pembelajaran konvensional kelas X SMA N 1 2X11 Enam Lingkung. Jenis penelitian yang dilakukan adalah penelitian eksperimen semu, dengan desain penelitian Randomized Control Group Only Design. Pengambilan sampel dilakukan dengan teknik Purposive Sampling. Data hasil belajar siswa diperoleh dari hasil tes akhir kedua kelas sampel yaitu hasil belajar setelah diterapkan pada kelas eksperimen yang menggunakan pendekatan berbasis kecerdasan dan hasil belajar pada kelas kontrol yang menggunakan pembelajaran konvensional. Secara umum hasil penelitian menunjukkan persentase ketuntasan ulangan akhir siswa sebesar 76,12% pada pembelajaran di kelas eksperimen. Dengan demikian dapat diketahui bahwa siswa kelas X1 SMA N 1 2X11 Enam Lingkung mempunyai rata-rata hasil belajar lebih besar atau sama dengan 75 dengan menggunakan pendekatan berbasis kecerdasan. Berdasarkan penelitian dapat disimpulkan bahwa hasil belajar matematika siswa yang diterapkan pendekatan berbasis kecerdasan lebih tinggi dibandingkan hasil belajar konvensional di SMA N 1 2X11 Enam Lingkung.

Multiphysics machine learning framework for on-demand multi-functional nano pattern design by light-controlled capillary force lithography

Nature finds ways to realize multi-functional surfaces by modulating nano-scale patterns on their surfaces, enjoying transparent, bactericidal, and/or anti-fogging features. Therein height distributions of nanopatterns play a key role. Recent advancements in nanotechnologies can reach that ability via chemical, mechanical, or optical fabrications. However, they require laborious complex procedures, prohibiting fast mass manufacturing. This paper presents a computational framework to help design multi-functional nano patterns by light. The framework behaves as a surrogate model for the inverse design of nano distributions. The framework’s hybrid (i.e., human and artificial) intelligence-based approach helps learn plausible rules of multi-physics processes behind the UV-controlled nano patterning and enriches training data sets. Then the framework’s inverse machine learning (ML) model can describe the required UV doses for the target heights of liquid in nano templates. Thereby, the framework can realize multiple functionalities including the desired nano-scale color, frictions, and bactericidal properties. Feasibility test results demonstrate the promising capability of the framework to realize the desired height distributions that can potentially enable multi-functional nano-scale surface properties. This computational framework will serve as a multi-physics surrogate model to help accelerate fast fabrications of nanopatterns with light and ML.