Research & Publications

Exploring the frontiers of machine learning, computer vision, and medical AI.

Medical AI2025

DIAE: Dynamic Input-Adaptive Ensemble

BSc Graduation Project: A novel framework (DIAE) achieving 97.05% accuracy in thyroid cancer classification by dynamically adjusting ensemble weights.

Ensemble LearningMedical AI

Deep Learning2025

Interactive Ensemble Architecture

Ongoing Research: Moving beyond late fusion with a novel architecture where models collaborate and exchange information during feature extraction.

Deep LearningInnovation

Machine Learning2025

Attention-Weighted Learnable KNN

Research journey discovering that Test-Time Augmentation (not attention) drives calibration improvements. Honest assessment of limitations and systematic experimentation across 6 iterations.

Machine LearningResearch

0%

Research & Publications

DIAE: Dynamic Input-Adaptive Ensemble

Interactive Ensemble Architecture

Attention-Weighted Learnable KNN

Research & Publications

DIAE: Dynamic Input-Adaptive Ensemble

Interactive Ensemble Architecture

Attention-Weighted Learnable KNN