MAKE MDPI

🔥 Hot Paper on Medical Imaging and Disease Diagnosis in MAKE

Medical Image Classifications Using Convolutional Neural Networks: A Survey of Current Methods and Statistical Modeling of the Literature

👥 Authors: Foziya Ahmed Mohammed, Kula Kekeba Tune, Beakal Gizachew Assefa, Marti Jett, and Seid Muhie

How are convolutional neural networks transforming medical image analysis? 🏥🤖

This comprehensive review explores CNN-based approaches for medical image classification, covering modern architectures, frameworks, activation functions, ensemble methods, hyperparameter optimization techniques, performance metrics, datasets, and preprocessing strategies used in medical imaging. 📊

Beyond reviewing current methodologies, the authors apply statistical modeling to identify emerging trends, research gaps, and future directions in the field. The study highlights the growing adoption of CNN and CNN-transformer hybrid architectures while emphasizing the need for improved explainability, robustness, and reproducibility in medical AI. 🔬📈

As part of our collection on Hot Papers in Medical Imaging and Disease Diagnosis, this work provides valuable insights for researchers working at the intersection of artificial intelligence and healthcare. 🩺

📖 Read more:
https://www.mdpi.com/2504-4990/6/1/33

📢 New paper published in MAKE

A Survey on Self-Supervised Learning in Cybersecurity: Network Intrusion and Malware Detection

👥 Authors: Josue Genaro Almaraz-Rivera, Jose Antonio Cantoral-Ceballos, and Juan Felipe Botero.

How can self-supervised learning strengthen the next generation of cybersecurity solutions? 🤖🛡️

This review paper provides a comprehensive survey of self-supervised learning (SSL) applications in cybersecurity, with a particular focus on network intrusion detection and malware detection. The study reviews advances from 2019 to 2025, compares contrastive learning and auxiliary pretext-task approaches, and identifies key opportunities for future research, including encrypted traffic analysis, resource-constrained devices, malware analysis, and tabular cybersecurity data.

As one of the first comprehensive surveys dedicated to self-supervised learning in cybersecurity, this work offers valuable insights into emerging research directions and the growing role of SSL in building more robust, scalable, and data-efficient cyber defense systems. 🚀🛡️

📖 Read more:
https://www.mdpi.com/2504-4990/8/5/121

🔥 Highly Cited Paper from MAKE

📝 Alzheimer’s Disease Detection Using Deep Learning on Neuroimaging: A Systematic Review

👥 Authors: Mohammed G. Alsubaie, Suhuai Luo, and Kamran Shaukat

Alzheimer’s disease remains a major global health challenge, creating a strong need for accurate and reliable diagnostic approaches.

This systematic review examines recent research on Alzheimer’s disease detection using deep learning and neuroimaging data. The paper reviews single- and multi-modality approaches, biomarkers, preprocessing techniques, and deep learning architectures, including CNNs, RNNs, graph neural networks, and autoencoders.

The study highlights both the promise and limitations of deep learning for Alzheimer’s disease detection, emphasizing the need for robust datasets, discriminative feature representations, and standardized benchmark platforms.

🔗 Read more: https://www.mdpi.com/2504-4990/6/1/24

📢 New paper published in MAKE

Dialectal Landscape: A Sequence-to-Sequence Approach to Japanese Dialect-to-Standard Normalization

👥 Authors: Kinga Lasek, Michal Ptaszynski, Fumito Masui, Mujahid Khalifah, and Juuso Eronen

Regional dialects continue to create communication gaps in Japanese, especially in contexts such as healthcare and emergency communication.

This paper proposes a text-to-text normalization method for converting eight Japanese dialects into standard Japanese using a fine-tuned mT5-small sequence-to-sequence architecture.

The study evaluates training strategies, preprocessing methods, and model performance, showing that character-level evaluation can better capture dialect-to-standard normalization quality in Japanese text.

🔗 Read more: https://www.mdpi.com/2504-4990/8/5/115

🔥 XAI Paper Highlight from Make

📝 Using Segmentation to Boost Classification Performance and Explainability in CapsNets

👥 Authors: Dominik Vranay, Maroš Hliboký, László Kovács, and Peter Sinčák

Explainability remains a key challenge in image classification, especially when using complex neural network architectures.

This paper introduces Combined-CapsNet, a novel approach that integrates segmentation masks as reconstruction targets within Capsule Neural Networks. By focusing on significant image regions, the method aims to improve both classification performance and model explainability.

The study shows that segmentation-based reconstruction can help produce clearer and more interpretable explanations while reducing model complexity, supporting the development of more transparent image classification systems.

🔗 Read more: https://www.mdpi.com/2504-4990/6/3/68

📢 New paper published in MAKE

📝 Algorithmic Insights into Human Irrationality: Machine Learning Approaches to Detecting Cognitive Biases and Motivated Reasoning

👥 Authors: Sarthak Pattnaik, Chhayank Jain, and Eugene Pinsky

This study applies machine learning methods to large-scale public opinion data to examine cognitive biases and motivated reasoning.

Drawing on dual-process theory and nudge architecture, the authors use hierarchical unsupervised clustering and supervised predictive models to detect patterns related to loss aversion, availability heuristic, and partisan motivated reasoning.

The findings offer insights into political cognition, digital engagement, and the ethical use of AI-augmented behavioral analysis in an era of affective polarization.

🔗 Read more: https://www.mdpi.com/2504-4990/8/4/98

🎉 We are delighted to share that Machine Learning and Knowledge Extraction has achieved a CiteScore of 12.7 in the 2025 release.

This places MAKE among the top 10% of journals in the Scopus category Engineering (miscellaneous).

The new score represents a +28% increase from last year’s CiteScore of 9.9, reflecting the growing impact and visibility of the research published in the journal.

We are grateful to our authors, reviewers, Guest Editors, Editorial Board Members, and readers for their valuable contributions and continued support. 👏

📊 Full details: https://bit.ly/4e25s0c

🔥 Highly Cited Paper from MAKE

📝 Why Do Tree Ensemble Approximators Not Outperform the Recursive-Rule eXtraction Algorithm?

👥 Authors: Soma Onishi, Masahiro Nishimura, Ryota Fujimura, and Yoichi Hayashi

Interpretability remains a major challenge in machine learning, especially when using complex tree ensemble models.

This paper investigates why tree ensemble approximators do not necessarily outperform the Recursive-Rule eXtraction algorithm in terms of interpretability. The study proposes a metric for comparing the interpretability of different rule sets and examines how rule-based explanations differ across decision tree-based, unordered, and decision list-based approaches.

The findings highlight the importance of evaluating not only the number of rules but also their practical interpretability, especially when developing transparent and explainable AI systems.

🔗 Read more: https://www.mdpi.com/2504-4990/6/1/31

📢 New paper published in MAKE

📝 Diffusion-Based Feature Denoising and Using NNMF for Robust Brain Tumor Classification

👥 Authors: Hiba Adil Al-kharsan and Robert Rajko

Robust brain tumor classification from MRI scans remains an important challenge, especially as deep learning models can be vulnerable to adversarial perturbations.

This paper proposes a framework combining Non-Negative Matrix Factorization, lightweight CNN-based classification, and diffusion-based feature denoising to improve robustness in brain tumor classification.

The study highlights how interpretable NNMF-based representations and diffusion-based defense techniques can support more reliable medical image classification under adversarial conditions.

🔗 Read more: https://www.mdpi.com/2504-4990/8/4/105

📢 Highly Cited Papers on AI in Healthcare

Machine Learning and Knowledge Extraction is pleased to highlight a selection of highly cited papers on Artificial Intelligence in Healthcare.

This collection features influential research on AI-driven healthcare applications, including medical diagnosis, clinical decision support, medical imaging, predictive modeling, explainable AI, and trustworthy machine learning.

We invite researchers working in machine learning, artificial intelligence, healthcare technologies, medical diagnostics, and knowledge extraction to explore these highly cited contributions to the field.

🔗 Read more: https://bit.ly/3QbVvFM