KAUST Academy
العربيةApply now

Layan Zaafarani

After KAUST Academy

Layan graduated from King Abdulaziz University in 2024 and attended the 2024 KAUST Academy summer AI program at Lady Margaret Hall, Oxford. As of October 2026, Layan works at the National Center for Inspection and Monitoring.

Summer projects

Artificial Intelligence Specialization · Cohort 2024

Accelerated Post-Disaster Recovery: An AI-Powered System for Damage Assessment and Reconstruction Prioritization

This project introduces an innovative AI-driven framework to revolutionize post-disaster recovery efforts. Leveraging UAV imagery and advanced AI techniques, our system rapidly assesses damage to buildings and infrastructure using an Attention U-Net model for image segmentation. Unsupervised learning then clusters affected areas based on damage severity, building importance, and distance, utilizing both segmentation results and Google Maps metadata. A reinforcement learning model, employing PPO and DQN algorithms, optimizes resource allocation and prioritizes reconstruction efforts, aiming to minimize total recovery time. This integrated approach promises faster, more accurate damage assessment and data-driven decision-making, significantly reducing the time needed to restore affected communities and setting a new standard for efficient disaster response strategies.

See it in the program report (PDF)

Artificial Intelligence Specialization · Cohort 2023

Addressing Bias: Through Ensemble Learning and Regularized Fine-Tuning

Addressing biases in AI models is crucial for ensuring fair and accurate predictions. However, obtaining large, unbiased datasets for training can be challenging. Our project proposes a comprehensive approach using multiple methods to remove bias in AI models, with only a small dataset and a potentially biased pretrained model. We train multiple models with the counter-bias of the pretrained model through data splitting, local training, and regularized fine-tuning, gaining potentially counter-biased models. Then, we employ ensemble learning for all models to reach unbiased predictions. To further accelerate the inference time of our ensemble model, we conclude our solution with knowledge distillation that results in a single unbiased neural network. We demonstrate the effectiveness of our approach through experiments on the CIFARI0O and HAMIOO00 datasets, showcasing promising results. This work contributes to the ongoing effort to create more unbiased and reliable AI models, even with limited data availability.

See it in the program report (PDF)

With KAUST Academy

All alumni

Want your entry updated or removed? Contact us.