A Researcher from Northern Technical University Publishes an Advanced Model in Learning and Artificial Intelligence.

Assistant Professor Mohammed Talal Ghazal, a faculty member at the Engineering Technical College/Mosul – Department of Biomedical Engineering Technologies, has successfully published a research paper with his research team in Scientific Reports, a journal published under the Nature Portfolio of the global publishing house Springer Nature. The journal is ranked in the first quartile (Q1) and holds an internationally recognized impact factor.
This achievement comes within the framework of the rapid development in the field of Machine Learning. The study is entitled:
“Uncertainty-weighted semi-supervised learning with dynamic entropy masking and Bhattacharyya-regularized loss”
The study presents an innovative framework in Semi-Supervised Learning aimed at maximizing the effective utilization of unlabeled data more efficiently than traditional methods through:
An uncertainty-weighting mechanism that prioritizes samples with higher informational value.
A Dynamic Entropy Masking technique to reduce the impact of inaccurate labels.
The adoption of a Bhattacharyya-regularized loss function to enhance prediction consistency and improve model stability.
The results demonstrated a noticeable improvement ranging between 3–5% in classification accuracy compared to leading competing models, particularly in environments suffering from limited or imbalanced labeled data.
This scientific publication reflects the advanced research standing of Northern Technical University and affirms its active presence on the global scientific research map.
We extend our best wishes to our university’s researchers for continued excellence and sustained scientific contribution.