I'm a PhD researcher at the University of Bonn working within the AID4Crops unit. My research sits at the intersection of Uncertainty Quantification and Explainable Machine Learning — with the goal of building more transparent and reliable AI systems for crop monitoring and precision agriculture. 🌱

Before Bonn, I explored deep learning for medical imaging during my MSc, and I love how connected science really is across domains.

• Investigating the relationship between model uncertainty and explainability in deep learning models for crop monitoring
• Extending conformal prediction frameworks to dense prediction tasks (segmentation) in precision agriculture
• Supervising MSc and BSc students on uncertainty sensitivity and anomaly detection projects

• 🌾 Precision Agriculture — AI-driven crop monitoring and decision support
• 🤔 Uncertainty Quantification — Conformal prediction, ensemble methods, calibration
• 🔍 Explainable ML — GradCAM, attention maps, integrated gradients, mechanistic interpretability
• 🏥 Medical Imaging — Diabetic retinopathy classification, surgical tool segmentation

Journals

• 📘 Farag et al. (2025). Enhancing Decision Support in Crop Production: Analyzing Conformal Prediction for Uncertainty Quantification. Computers and Electronics in Agriculture (IF: 8.9, Q1)
• 📘 Emam, Farag & Roscher. Confident Naturalness Explanation (CNE). IEEE Geoscience and Remote Sensing Letters (IF: 4.0, Q1)
• 📘 Farag, Fouad & Abdel-Hamid. Automatic Severity Classification of Diabetic Retinopathy Based on DenseNet and CBAM. IEEE Access (IF: 3.6)

Workshops

• 🔧 Emam, Farag et al. (2025). A Framework for Enhanced Decision Support in Digital Agriculture Using Explainable ML. ECCV 2024 Workshops, Springer LNCS
• 🔧 Farag, Kierdorf & Roscher (2023). Inductive Conformal Prediction for Harvest-Readiness Classification of Cauliflower Plants. ICCV 2023 Workshops

📚 Full list: Google Scholar · ResearchGate

📈 Research Impact (update periodically)

Citations	h-index	Publications
	—	5 (journals + workshops)

💡 Get your live numbers from Google Scholar and update the table above.

Deep Learning & ML

Architectures I've worked with: CNNs, ViTs, UNet, TransUNet, DeepLab, Mask2Former, YOLOS

Self-supervised learning: DINO, SwAV, BarlowTwins, BYOL, SimCLR, SimSiam, NNCLR

I enjoy mentoring — it's one of my favourite parts of academia.

Courses (TA @ University of Bonn)

• Explainable Machine Learning
• Acquisition, Analysis and Modelling of Phenotypes
• Advanced Methods for Heterogeneity and Phenotype Analysis

Student Supervision

• 🎓 3 MSc students — topics spanning UQ, anomaly detection, and data dimensionality (with Fraunhofer, DLR, and Erasmus programme)
• 🎓 2 BSc students — sensitivity of model uncertainty to hyperparameter variations (with GUC)

• 🗣️ Fluent in English and Arabic
• ⚽ Big football fan · 🎞️ Video editing enthusiast · 🎮 Occasional gamer
• 🧠 Strong believer that science is interconnected — from crops to the clinic

I'm always happy to connect with researchers and practitioners working on:

• 🌾 AI for precision agriculture or remote sensing
• 🤔 Uncertainty quantification in deep learning
• 🔍 Explainability methods for computer vision
• 🏥 Medical image analysis

If any of this overlaps with your work — feel free to reach out!