Explainable AI: Bridging the Gap between Machine Learning Models and Human Understanding

Journal of Informatics Education and Research, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 1, 2024

Explainable AI (XAI) is one of the key game-changing features in machine learning models, which contribute to making them more transparent, regulated and usable different applications. In (the) investigation this paper, we consider four rows explanation methods—LIME, SHAP, Anchor, Decision Tree-based Explanation—in disentangling decision-making process black box models within fields. our experiments, use datasets that cover domains, for example, health, finance image classification, compare accuracy, fidelity, coverage, precision human satisfaction each method. Our work shows rule trees approach called (Decision explanation) mostly superior comparison other non-model-specific methods performing higher coverage regardless classifier. addition this, respondents who answered qualitative evaluation indicated they were very content with decision tree-based explanations these types are easy understandable. Furthermore, most famous sorts clarifications instinctive significant. The over discoveries stretch on utilize interpretable strategies facilitating hole between understanding thus advancing straightforwardness responsibility AI-driven decision-making.

Language: Английский

---

Infection and Inflammation in Nuclear Medicine Imaging: The Role of Artificial Intelligence

Seminars in Nuclear Medicine, Journal Year: 2025, Volume and Issue: unknown

Published: March 1, 2025

Language: Английский

---

The Use of Hybrid CNN-RNN Deep Learning Models to Discriminate Tumor Tissue in Dynamic Breast Thermography

Journal of Imaging, Journal Year: 2024, Volume and Issue: 10(12), P. 329 - 329

Published: Dec. 21, 2024

Breast cancer is one of the leading causes death for women worldwide, and early detection can help reduce rate. Infrared thermography has gained popularity as a non-invasive rapid method detecting this pathology be further enhanced by applying neural networks to extract spatial even temporal data derived from breast thermographic images if they are acquired sequentially. In study, we evaluated hybrid convolutional-recurrent network (CNN-RNN) models based on five state-of-the-art pre-trained CNN architectures coupled with three RNNs discern tumor abnormalities in dynamic images. The architecture that achieved best performance was VGG16-LSTM, which showed accuracy (ACC), sensitivity (SENS), specificity (SPEC) 95.72%, 92.76%, 98.68%, respectively, CPU runtime 3.9 s. However, fastest AlexNet-RNN 0.61 s, although lower (ACC: 80.59%, SENS: 68.52%, SPEC: 92.76%), but still superior AlexNet 69.41%, 52.63%, 86.18%) 0.44 Our findings show CNN-RNN outperform stand-alone models, indicating recovery thermographs possible without significantly compromising classifier runtime.

Language: Английский

---

Information Geometry and Manifold Learning: A Novel Framework for Analyzing Alzheimer’s Disease MRI Data

Diagnostics, Journal Year: 2025, Volume and Issue: 15(2), P. 153 - 153

Published: Jan. 10, 2025

Background: Alzheimer’s disease is a progressive neurological condition marked by decline in cognitive abilities. Early diagnosis crucial but challenging due to overlapping symptoms among impairment stages, necessitating non-invasive, reliable diagnostic tools. Methods: We applied information geometry and manifold learning analyze grayscale MRI scans classified into No Impairment, Very Mild, Moderate Impairment. Preprocessed images were reduced via Principal Component Analysis (retaining 95% variance) converted statistical manifolds using estimated mean vectors covariance matrices. Geodesic distances, computed with the Fisher Information metric, quantified class differences. Graph Neural Networks, including Convolutional Networks (GCN), Attention (GAT), GraphSAGE, utilized categorize levels graph-based representations of data. Results: Significant differences structures observed, increased variability stronger feature correlations at higher levels. distances between Impairment Mild (58.68, p<0.001) (58.28, are statistically significant. GCN GraphSAGE achieve perfect classification accuracy (precision, recall, F1-Score: 1.0), correctly identifying all instances across classes. GAT attains an overall 59.61%, variable performance Conclusions: Integrating geometry, learning, GNNs effectively differentiates AD stages from The strong indicates their potential assist clinicians early identification tracking progression.

Language: Английский

---

Detection of Architectural Dysplastic Features from Histopathological Imagery of Oral Mucosa Using Neural Networks

Bioengineering, Journal Year: 2025, Volume and Issue: 12(3), P. 216 - 216

Published: Feb. 20, 2025

Oral cancer is a serious illness, but it potentially curable if early detection can be achieved successfully. epithelial dysplasia (OED), which precursor to oral squamous cell carcinoma (OSCC), provide abnormal characteristics diagnose the risk of developing cancer. This paper proposes neural network architecture for detecting dysplastic features architecture, including irregular stratification and bulbous rete ridges. The different combinations atrous convolution, batch normalization, global pooling, dropout are discussed regarding their effects, along with an ablation study. A signature library containing image patches was constructed utilized train models. best-performing model in validation set attained average accuracy 97.52%. results blind test from receiver operating characteristic (ROC) curves show that best reached probability detection, 0.8571, stratifications 0.8462

Language: Английский

---

Artificial Intelligence With Neural Network Algorithms in Pediatric Astrocytoma Diagnosis: A Systematic Review

Innovative medicine of Kuban, Journal Year: 2025, Volume and Issue: 10(1), P. 93 - 100

Published: Feb. 26, 2025

Background: Astrocytoma is a common pediatric brain tumor that poses significant health burden. Recent advancements in artificial intelligence (AI), particularly neural network algorithms, have been studied for their precision and efficiency medical diagnostics via effectively analyzing imaging data to identify patterns anomalies. Objective: To systematically review AI-based diagnostic tools with algorithms’ methodologies, sensitivities, specificities, potential clinical integration astrocytoma, providing consolidated perspective on overall performance impact decision-making. Methods: As per PRISMA 2020 guidelines, we conducted comprehensive search PubMed, Scopus, ScienceDirect February 5, 2024. The strategy was guided by PECO question focusing astrocytoma diagnosis using AI algorithms vs computed tomography or magnetic resonance (MRI). Keywords were terms related algorithms. We included studies the accuracy of methods cases (World Health Organization grades 1-3), no restrictions publication year country. excluded papers written languages other than English Bahasa Indonesia nonhuman studies. Data assessed Effective Public Practice Project tool. Results: Of 454 articles screened, 6 met inclusion criteria. These varied design, location, sample size, ranging from 10 135 subjects. showed high sensitivity specificity, often surpassing traditional radiological techniques. Notably, 3-dimensional MRI demonstrated improved compared 2-dimensional (96% 77%). models exhibited levels comparable exceeding expert radiologists, metrics such as classification 92% values area under receiver operating characteristic curve. Conclusions: shows promise enhancing diagnosis. reviewed indicate these advanced can achieve superior specificity conventional Integrating into practice could substantially improve patient outcomes.

Language: Английский

---

DESENVOLVIMENTO DE UM MODELO DE MACHINE LEARNING PARA CLASSIFICAÇÃO DE IMAGENS DE RAIO-X DO TÓRAX EM DOENÇAS RESPIRATÓRIAS

Revista Foco, Journal Year: 2025, Volume and Issue: 18(3), P. e8110 - e8110

Published: March 31, 2025

As doenças respiratórias representam uma das principais causas de morbidade e mortalidade global, destacando a necessidade diagnósticos rápidos precisos. Este estudo propõe o desenvolvimento modelos machine learning utilizando redes neurais convolucionais (CNNs) para classificar imagens raio-X do tórax em quatro categorias: COVID-19, normal, pneumonia viral bacteriana. Utilizando as arquiteturas ResNet50 YOLOv8 pré-treinadas técnica transferência aprendizagem, os foram adaptados contexto específico radiografias pulmonares. O treinamento foi realizado com conjuntos dados balanceados diversas técnicas pré-processamento aumento aplicadas. Os resultados indicam um desempenho promissor ambos modelos, elevada acurácia na classificação diferentes patologias, demonstrando potencial da abordagem auxiliar profissionais saúde ambientes recursos limitados, como Unidade Pronto Atendimento (UPA) Ariquemes-RO, podendo ser estendida outras unidades saúde.

---

Challenges, optimization strategies, and future horizons of advanced deep learning approaches for brain lesion segmentation

Methods, Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Language: Английский

---

Metric-Based Meta-Learning Approach for Few-Shot Classification of Brain Tumors Using Magnetic Resonance Images

Electronics, Journal Year: 2025, Volume and Issue: 14(9), P. 1863 - 1863

Published: May 2, 2025

Brain tumor prediction from magnetic resonance images is an important problem, but it difficult due to the complexity of brain structure and variability in appearance. There have been various ML DL-based approaches, limitations current models are a lack adaptability new tasks need for extensive training on large datasets. To address these issues, novel meta-learning approach has proposed, enabling rapid adaptation with limited data. This paper presents method that integrates vision transformer metric-based model, few-shot learning enhance classification performance. The proposed begins preprocessing MRI images, followed by feature extraction using transformer. A Siamese network enhances model’s learning, quick unseen data improving robustness. Furthermore, applying strategy performance when there comparison other developed reveals consistently performs better. It also compared previously approaches same datasets evaluation metrics including accuracy, precision, specificity, recall, F1-score. results demonstrate efficacy our methodology classification, which significant implications enhancing diagnostic accuracy patient outcomes.

Language: Английский

---

Image Text Extraction and Natural Language Processing of Unstructured Data from Medical Reports

Machine Learning and Knowledge Extraction, Journal Year: 2024, Volume and Issue: 6(2), P. 1361 - 1377

Published: June 18, 2024

This study presents an integrated approach for automatically extracting and structuring information from medical reports, captured as scanned documents or photographs, through a combination of image recognition natural language processing (NLP) techniques like named entity (NER). The primary aim was to develop adaptive model efficient text extraction report images. involved utilizing genetic algorithm (GA) fine-tune optical character (OCR) hyperparameters, ensuring maximal length, followed by NER categorize the extracted into required entities, adjusting parameters if entities were not correctly based on manual annotations. Despite diverse formats images in dataset, all Russian, this serves conceptual example (IE) that can be easily extended other languages.

Language: Английский

---