Economic geology & mineral geochemistry
Sulfide mineral chemistry, trace elements, S-Pb isotopes, critical metals, ore-system discrimination, metallogenic interpretation, and mineral-system reasoning.
Economic Geology · GeoAI · Mineral Discovery
Turning geoscience data into defensible exploration decisions.
I am Dr. Muhammad Amar Gul, a geoscientist and applied AI/ML researcher connecting mineral geochemistry, mineral systems, spatial data science, remote sensing, uncertainty, and scientific software for mineral exploration.
10+ yearsGeoscience experience
20+Peer-reviewed publications
138 depositsGlobal pyrite research coverage
37 depositsGalena metallogenic AI study
Research identity
Economic geology and AI, connected through geological meaning.
My work combines domain geoscience with machine learning, spatial analysis, uncertainty, and reproducible software. Prediction is useful only when the data authority, validation design, transfer domain, and geological interpretation are defensible.
GeoAI & scientific machine learning
Grouped validation, interpretable ML/DL, anomaly detection, classification, uncertainty-aware prediction, SHAP/LIME, clustering, and reproducible evaluation.
Spatial exploration & remote sensing
Mineral prospectivity, alteration mapping, geological and structural evidence integration, geophysics, terrain, geostatistics, GIS, and spatial validation.
Scientific software & data systems
Python research pipelines, ETL, QA gates, Streamlit/Gradio tools, automation, batch execution, evidence inventories, and auditable technical delivery.
Research questions, programs, methods, and validation principles →
Featured research
From global mineral chemistry to interpretable metallogenic models.
Selected peer-reviewed work links geochemical data, geological context, machine learning, and explicit validation rather than treating model accuracy as the scientific endpoint.
Global pyrite metallogenic typing
~5,200 analyses from 138 global deposits/settings, multielement ML, class-balance testing, blind evaluation, and deposit-scale LOGO validation.
Paper → Repository →Galena geochemistry & metallogenic discrimination
Galena from 37 Pb-Zn deposits evaluated with RF, GB, MLP, SVM, imbalance treatment, feature importance, and high-dimensional visualization.
Paper → Repository →Gunga pyrite, isotopes & machine learning
Pyrite trace elements and isotopic signatures integrated with global comparison data and deposit-aware validation for Pb-Zn metallogenic interpretation.
Paper → Repository →Gunga sphalerite, deep learning & critical metals
Sphalerite geochemistry and S-Pb isotopes combined with deep learning to investigate ore genesis and critical-metal enrichment, including Ge-bearing sphalerite.
Paper → Repository →Publication status, broader geoscience research, and citation links →
Selected systems & professional R&D
Research translated into exploration workflows.
These systems span research and professional R&D. Public descriptions focus on scientific and technical capability; proprietary employer or client code, data, thresholds, and targets are not published without authorization.
GeoAnomalyAI
Multivariate anomaly intelligence using correlation structure, robust outlier diagnostics, clustering, anomaly scoring, spatial confidence, threshold sensitivity, and target ranking.
GeoSpectraAI
Remote-sensing/spectral, geological, geochemical, structural, field, and spatial evidence integrated for prospectivity screening and confidence assessment.
Target detection
Target ranking, uncertainty products, geostatistical outputs, spatial validation, map QA, batch execution, and reproducible field-oriented delivery.
Geostatistical & GPU R&D
Variography, kriging, sequential Gaussian simulation, uncertainty analysis, Python implementation, and CUDA/GPU experimentation.
Project scope, methods, public-release boundaries, and exploration applications →
Current roles
Research, exploration delivery, and GeoAI development.
Senior Geoscientist (Data & AI)
Erity Pty Ltd. · Remote, Australia
Project Geologist (AI & ML)
China National Geological & Mining Corporation · Saudi Arabia
AIMEX Lab
Independent research initiative for artificial intelligence in mineral exploration and reproducible GeoAI benchmarking.
How I work
Prediction is a computational result; geological interpretation is a scientific argument.
Data authority first
Provenance, QA/QC, exclusions, analytical comparability, and geological labels are defined before modelling.
Validate at the right geological scale
Grouped, deposit-aware, spatial, geological, external, or prospective validation is preferred where random row splitting would leak context.
Represent uncertainty
Threshold sensitivity, cohort choices, spatial uncertainty, transferability, and failure modes are treated as part of the result.
Separate prediction from mechanism
Feature importance and high model accuracy can support investigation, but do not by themselves establish geological causality.
AIMEX
Research initiative
AIMEX Lab — an independent GeoAI research initiative.
I founded AIMEX Lab as a platform for collaborative GeoAI research, reproducible benchmarking, and mineral-exploration methodology. AIMEX has its own website and GitHub organization and complements my personal research portfolio.
Visit AIMEX LabCollaboration
Research, GeoAI, and mineral-exploration collaboration.
For research collaboration, technical benchmarking, scientific discussion, selected applied work, or relevant professional opportunities, contact me with the problem, available data, geological context, and intended decision.