On January 15, 2025, Jasmin Gründling-Riener, TU Wien Vice-Rector for Academic Affairs, and Daniel Löcker from the city’s Department of Cultural Affairs presented the Dissertation Award 2024 of the City of Vienna at TU Wien. Muhammad Abdullah Hanif won the Award, endowed with 1000 Euros, for his thesis “Robust and Energy Efficient Deep Learning Systems; he was supervised by Muhammad Shafique.

In his thesis, Muhammad Abdullah Hanif focuses on improving the reliability and energy efficiency of systems that use Deep Neural Networks (DNNs). Deep Learning (DL) algorithms have become essential in Artificial Intelligence (AI) due to their ability to uncover patterns in data. Ensuring their reliability against hardware-induced faults (like soft errors, aging, or manufacturing defects) is critical for domains in which safety is essential - like autonomous driving and healthcare. The dissertation explores low-cost, energy-efficient strategies to address hardware reliability issues, including the exploration of techniques that mitigate hardware faults. Key strategies to mitigate hardware faults developed in the thesis are protecting the most critical parts of a DNN that have the greatest impact on accuracy, extending the lifespan of hardware components storing DNN data, and introducing small, controlled errors in less important parts of the computation, saving energy without reducing accuracy. With these approaches, the dissertation combines software and hardware-level innovations to enhance DNN robustness and energy efficiency, achieving significant reliability improvements and efficiency gains while maintaining application-level accuracy.

Congratulations to Muhammad Abdullah Hanif for his outstanding scientific achievement!

Curious about our other news? Subscribe to our news feed, calendar, or newsletter, or follow us on social media.