Doctoral Researcher · AI & Clinical LLM
Aleka Melese
Ayalew
Center for Machine Vision and Signal Analysis (CMVS), University of Oulu, Finland
Doctoral Researcher · AI & Clinical LLM
Center for Machine Vision and Signal Analysis (CMVS), University of Oulu, Finland
I am a doctoral researcher at the Center for Machine Vision and Signal Analysis (CMVS), Faculty of Information Technology and Electrical Engineering, University of Oulu, Finland. My research focuses on applying artificial intelligence and agentic AI to real-world healthcare challenges, particularly in medical image analysis, disease detection, and large language models for clinical decision support.
My work spans a broad spectrum of healthcare AI: from X-ray-based detection of cardiopulmonary conditions (cardiomegaly, pneumonia, atelectasis) to smart IoT-integrated diagnostics for infectious diseases such as malaria and cancer screening. I leverage state-of-the-art deep learning architectures, CNNs, DenseNet-169, ResNet-50, Inception, and Involutional Neural Networks combined with advanced image processing techniques to build clinically relevant AI systems.
Before joining the University of Oulu, I spent six years at a university lecturer and educational quality assurance coordinator in Ethiopia, contributing to teaching and research in computer science and information technology. I am also an active peer reviewer, having evaluated 60+ manuscripts across 20+ international journals, conferences, and workshops, including NeurIPS events.
Center for Machine Vision and Signal Analysis (CMVS), University of Oulu, Finland. Aleka.Ayalew@oulu.fi
X-ray-based detection of cardiomegaly, pneumonia, and atelectasis; IoT-integrated cancer and infectious disease diagnostics.
Reviewed 60+ articles across 20+ journals, conferences, and workshops including NeurIPS-organised events.
Articles in Intelligent Systems with Applications, Sensing and Imaging, Multimedia Tools and Applications, and more.
Updates on publications, research milestones, and academic activities.
PEFT: Evaluating parameter-efficient fine-tuning for clinical-note summarization across transformers and large language models
Our Paper is published in Information Processing & Management (DOI: https://doi.org/10.1016/j.ipm.2026.105158), authored by Aleka Melese Ayalew, Md Rabiul Hasan, Tapio Seppänen, and Mourad Oussalah.
Large Language Models for Explainable Medical Text Summarization
Our systematic literature review is published in WIREs Data Mining and Knowledge Discovery (DOI: 10.1002/widm.70089), authored by Aleka Melese Ayalew, Md Rabiul Hasan, Tapio Seppänen, and Mourad Oussalah.
Preventing Content Leakage in LLM-Based Medical RAG:Structure-Only Retrieval for Faithful Clinical Summarization
Published as a SIGIR-ICTIR conference (DOI: 10.1145/3805713.3820416), authored by Aleka Melese Ayalew, Tapio Seppänen, and Mourad Oussalah.
Structured LLM-Generated Clinical Discharge Summaries for Nordic Healthcare Digitalization and Wireless Health Integration
Published as a book chapter (DOI: 10.1007/978-3-032-28819-6_38), authored by Aleka Melese Ayalew, Md Rabiul Hasan, Tapio Seppänen, and Mourad Oussalah.
Beyond Lexical Similarity: Evaluating Faithfulness in LLM-Based Medical Question Reformulation
Published as a Workshop (DOI: 10.18653/v1/2026.smm4h-1.16), authored by Md Rabiul Hasan, Aleka Melese Ayalew, and Mourad Oussalah.
MEG-QA: Medical Evidence-Grounded Question Abbreviation for Consumer Health Queries via Retrieval-Augmented Generation and Knowledge Graph Constraints in Digital Health
Published as a book chapter (DOI: 10.1007/978-3-032-28819-6_44), authored by Md Rabiul Hasan, Aleka Melese Ayalew, and Mourad Oussalah.
An explainable hybrid deep learning system for tuberculosis detection with Grad-CAM
Published in Discover Computing (DOI: 10.1007/s10791-025-09791-z), authored by Aleka Melese Ayalew, Nigus Wereta Asnake, Getnet Demil, and Mourad Oussalah.
Smart Breast Cancer Detection via Involutional Neural Networks
Early breast cancer diagnosis by transitioning from CNNs to Involutional Neural Networks, integrated with smart wearable devices for continuous remote monitoring.
Two-Phase Transfer Learning for GI Disease Classification
Novel two-phase transfer learning framework for automated gastrointestinal disease classification from endoscopy video data — improving clinical accuracy and accessibility.
Cardiomegaly Detection from X-rays — Intelligent Systems with Applications
ResNet-50 achieving 99.8% test accuracy for early-stage cardiomegaly detection published in Intelligent Systems with Applications.
Smart Malaria Classification with IoT-Based ML — Sensing and Imaging
IoT-integrated machine learning pipeline for early malaria detection. SVM model achieving 96% test accuracy and 95% AUC published in Sensing and Imaging.
Bridging artificial intelligence and real-world healthcare through medical image analysis, deep learning, and smart diagnostic systems.
Applying large language models in clinical settings: automating administrative data processing, reducing documentation workload, and generating real-time predictive analytics to support medical decision-making and keep practitioners informed of patient trends and medical literature.
Deep learning detection and classification of diseases from chest X-rays: cardiomegaly (99.8% accuracy via ResNet-50), pneumonia (YOLOv3 + SVM pipeline), atelectasis (CNNs with Anisotropic Diffusion filtering), and GI diseases from endoscopy videos using two-phase transfer learning.
AI-powered disease monitoring systems integrating wearable sensors, edge computing, and cloud ML models. Malaria classification (SVM/ANN/KNN with 96% test accuracy), breast cancer screening with Involutional Neural Networks, and real-time predictive analytics for clinical deployment.
Autonomous and human-in-the-loop AI agents that synthesize multimodal clinical data, retrieve evidence, and support timely, transparent, and context-aware care decisions.
Multimodal fine-tuning LLMs for integrated text, image, and clinical data understanding in healthcare AI applications.
LLMs and NLP systems for clinical text, administrative automation, evidence synthesis, and medical documentation.
Deep learning for disease detection and classification from X-ray, CT, and endoscopy images in clinical settings.
From teaching and research in Ethiopia to cutting-edge AI research at one of Finland's leading machine vision laboratories.
Doctoral research in deep learning and computer vision for medical image analysis. Current focus includes LLM applications in healthcare, AI-based disease detection from medical images, and smart IoT-integrated diagnostic systems. 24+ peer-reviewed publications during the programme.
Current PositionSix years of combined teaching and research in computer science and information technology. Responsibilities included curriculum development, quality assurance coordination, and applied research in machine learning, IoT-based systems, and health informatics.
Graduate studies in information technology, laying the foundation for AI and machine learning research. Early coursework in software engineering, algorithms, data structures, and emerging data-driven technologies.
30+ publications · Full list on Google Scholar and ORCID.
Information Processing & Management, 2026-06
DOI →WIREs Data Mining and Knowledge Discovery, 2026-06
DOI →Conference, 2026
SIGIR - ICTIR conference, 2026
DOI →book chapter, 2026
DOI →Discover Computing, 2025-11-12
DOI →Intelligent Systems with Applications, 2024
Full Article →Sensing and Imaging, 2024
DOI →2024 Second International Conference Computational and Characterization Techniques in Engineering & Sciences (IC3TES), 2024-11-15
DOI →International Conference on Decision Aid Sciences and Applications, 2022 · WOSUID: WOS:000839386600116
DOI →Medical imaging and deep learning
No publications match your search. Try different keywords.
Open to research collaborations, academic discussions, and opportunities in AI, medical imaging, and healthcare technology.
Aleka.Ayalew@oulu.fi — institutional address at University of Oulu.