Founder & Director — AIMEX Lab
Artificial Intelligence for Mineral Exploration Laboratory

Dr. Muhammad Amar Gul is a Research Scientist working at the interface of AI4Science and Earth-system processes, with a specialization in scientific machine learning for mineral systems. His research focuses on developing interpretable, uncertainty-aware, and data-efficient machine learning methodologies tailored for limited, imbalanced, and heterogeneous geoscientific datasets.

With more than a decade of experience spanning academia and industry across Saudi Arabia, Australia, and China, he integrates ensemble learning, deep neural networks, explainable AI (SHAP, attribution analysis), and probabilistic modeling to advance metallogenic discrimination, tectonic-environment prediction, and large-scale geochemical data mining.

As Founder and Director of AIMEX Lab (Artificial Intelligence for Mineral Exploration Laboratory), he leads the development of robust AI frameworks that bridge advanced computational methodologies with real-world mineral exploration challenges, contributing to sustainable critical-mineral resource development and next-generation AI-driven discovery.

I pursue methodology-driven machine learning for scientific discovery (AI4Science), focusing on probabilistic, uncertainty-aware, interpretable, and robust ML for limited, imbalanced, and heterogeneous datasets.

Core Interests

• Robust Generalization under data scarcity
• Imbalanced & heterogeneous dataset learning
• Interpretable AI (SHAP, feature attribution)
• Out-of-distribution validation
• AI for Critical Mineral Systems

Bridging AI theory and mineral system science for sustainable resource development.

• Senior AI-Research Scientist (Geoscience & Mining Applications) — Erity Pty Ltd., Perth, Australia (2026–Present)
• Applied Machine Learning Researcher (Geoscience & Mining Applications) — China National Geological & Mining Corporation, Saudi Arabia (2024–Present)

• Robust multi-class learning under severe class imbalance and distribution shift
• Large-scale generalization across 5,000+ samples from 100+ independent sources
• Interpretable scientific ML with SHAP-based diagnostics and stability analysis
• Reproducible research tooling using Python + interactive apps for benchmarking and deployment readiness

• Built CNN, Random Forest, SVM, Gradient Boosting, XGBoost, and MLP pipelines for multi-class classification
• Studied effects of class imbalance on uncertainty, stability, and generalization
• Applied SHAP for feature diagnostics and model interpretation

• Curated 5,000+ samples from 100+ independent sources
• Evaluated out-of-distribution performance across distinct data-generating processes
• Demonstrated strong generalization with AUC > 0.99 across independent test sets

• Tested model behavior under noise, sparsity, and missing data
• Integrated uncertainty-aware evaluation into ML workflows
• Built interactive research tools using Python + Gradio

• Big Data mining on Galena geochemistry using machine learning algorithms: Implications for Metallogenic Discrimination — Mathematical Geosciences

• Artificial Intelligence-Driven Metallogenic Typing of Pyrite from Global Ore Systems — Journal of Geochemical Exploration

• Machine Learning-Driven Classification of Pb–Zn Ore Deposits Using Pyrite Trace Elements and Isotopic Signatures: A Case Study of the Gunga Deposit — Journal of Geochemical Exploration

Random Forest | XGBoost | CNN | SVM | MLP | Gradient Boost

Cross-Validation | Leave-One-Group-Out | SHAP | t-SNE | UMAP
Class Imbalance Handling | SMOTE | RUC

Geochemical Big Data | LA-ICP-MS | Isotope Geochemistry
GIS | Remote Sensing | ETL | ISO/OGC Data Modeling

AIMEX Lab develops interpretable and robust AI systems for mineral exploration and scientific geoscience, bridging advanced ML methodology with real-world Earth system datasets.

AI × Critical Minerals × Interpretable Science × Robust ML