MAKE MDPI

📢 Call for Papers | Special Issue in Machine Learning and Knowledge Extraction

📝 Knowledge Discovery and Extraction with Large Language Models, Retrieval-Augmented Generation, and Autonomous Agents: Methods, Architectures, and Applications

Machine Learning and Knowledge Extraction (MAKE) is pleased to invite submissions to this Special Issue, which focuses on emerging methods and advanced approaches for knowledge discovery and extraction using Large Language Models, Retrieval-Augmented Generation, and autonomous AI agents.

This Special Issue welcomes contributions on LLM-based knowledge extraction, RAG architectures, autonomous agents, information retrieval, knowledge representation, AI-driven decision support, and real-world applications across scientific, industrial, and data-intensive domains.

👥 Guest Editors: Dr. Shaheen Khatoon, Azhar Mahmood, and Marek Sikora

📅 Submission deadline: 31 December 2026

🔗 Learn more and submit your manuscript: https://www.mdpi.com/journal/make/special_issues/IHZ71Z7Y03

🔥 Highly Cited Paper from MAKE

📝 Cross-Validation Visualized: A Narrative Guide to Advanced Methods

👥 Authors: Johannes Allgaier and Rüdiger Pryss

Choosing the right cross-validation method is essential for reliable machine learning model evaluation, but the wide range of available techniques can make this decision challenging.

This paper provides a clear narrative and visual guide to advanced cross-validation methods, including leave-one-out, leave-p-out, Monte Carlo, grouped, stratified, time-based, and block cross-validation approaches.
By clarifying key terminology and presenting graphical explanations, the study offers practical guidance for both early-career and experienced researchers working with machine learning model selection and performance reporting.

🔗 Read more: https://www.mdpi.com/2504-4990/6/2/65

📢 New paper published in MAKE

📝 Explainable Combined Spatial Representations for ECG Arrhythmia Classification

👥 Authors: Iulia Onică and Iulian Ciocoiu

The paper introduces a novel input fusion strategy for ECG arrhythmia classification by combining various spatial representations of time-series recordings. The approach utilizes four distinct transformation methods—spectrograms, Gramian Angular Fields (GAF), Recurrence Plots (RP), and the S-Transform (ST)—to generate composite images from single-lead ECG data. The fused 2x2 images are classified using both custom deep learning models and the ResNet-50 architecture. To ensure model transparency, several explainability algorithms were employed to highlight the specific image regions driving the classification. Benchmarked against the MIT-BIH and Chapman–Shaoxing datasets, the proposed method achieved high-performance metrics, including 99% accuracy, a 98.6% F1-score, and an AUC of 0.999, rivaling more complex state-of-the-art approaches.

🔗 Read the full paper: https://www.mdpi.com/2504-4990/8/5/114

🚀 First Paper Published in a MAKE Special Issue

We are pleased to announce the first publication in the Special Issue
“Next-Generation TinyML: Innovations in Models, Security, and Applications for Constrained Intelligent Systems.”

📄 Featured paper:
“A Tiny Vision-Based Model for Real-Time Student Attention Detection in Online Classes”

👥 Authors: Chaymae Yahyati, Ismail Lamaakal, Yassine Maleh, Khalid El Makkaoui , PhD, SMIEEE and ibrahim ouahbi

This paper presents a TinyML vision-based framework for real-time student attention detection in online learning environments, combining CNNs, BiLSTMs, and lightweight GRUs to estimate four attention levels directly from we**am clips. 🎓🤖

The proposed approach achieves strong performance on the DAiSEE benchmark while maintaining low computational overhead through structured pruning, INT8 quantization, and ONNX deployment, enabling efficient and privacy-preserving on-device analytics. 📊

This Special Issue aims to showcase advances in TinyML, efficient AI systems, edge intelligence, and secure intelligent applications for resource-constrained environments. 🚀

📖 Read the paper: https://www.mdpi.com/2504-4990/8/5/116

📌 Explore the Special Issue: https://www.mdpi.com/journal/make/special_issues/2W39J4533Q