Master’s Thesis at the University of Ninevah Explores Approximate Computing for Lightweight Neural Networks

Researcher Rahma Al-Rahman Wafeer Abdul-Ghani Al-Salih defended her Master’s thesis on Tuesday, 8 September 2026, at the College of Electronics Engineering, University of Ninevah. The thesis was entitled:

“Approximate Computing for Lightweight Neural Networks”

The thesis aimed to develop a framework for enhancing the efficiency of artificial neural network accelerators by employing approximate computing and genetic algorithms, thereby contributing to reducing energy consumption and hardware area while maintaining high performance accuracy.

The results demonstrated a reduction in energy consumption of up to 36.46% and a reduction in hardware area of approximately 15%, while maintaining high levels of classification accuracy. These findings support the development of FPGA-based artificial intelligence accelerators.

The examination committee consisted of:

  • Professor Dr. Abdul-Sattar Mohammed Khudr — Chair.
  • Assistant Professor Dr. Mohammed Abdul-Mutallab Mohammed — Member.
  • Lecturer Dr. Mamoun Abdul-Jabbar Dhunoon — Member.
  • Lecturer Dr. Imad Atiya Khalaf — Member and Supervisor.

At the conclusion of the defense, the committee commended the researcher’s scientific efforts and the thesis findings in the field of low-power computing and the development of artificial intelligence accelerators.