AI in Mineral Exploration: Methods, Data Sources, and Challenges
AI is changing how mineral exploration targets get generated and ranked. How AI-powered methods compare to traditional ones, what data they run on, and where they still fall short.
All
AI 70
Geology 26
Benchmarking 24
Machine Learning 17
Evaluation 12
Agentic AI 10
Open Source 9
Reinforcement Learning 5
China 4
Complexity Theory 3
Geology
Trust Boundaries for a Local MCP Server That Reads Your Files
MCP server security for a local server that reads your files: where Stratigraphic Amenity draws its trust boundaries, and what it leaves to the host and OS.
How to Find the UTM Zone EPSG Code of a Scanned Map
A UTM zone EPSG code is a simple sum once datum and hemisphere are known. How we turn map text like "UTM N83 Zone 15" into EPSG:26915, and where it fails.
Keeping Large Datasets Out of an AI Agent's Context Window
An AI agent context window fills fast when a tool reads a whole dataset. How our harness keeps large tables on disk and passes paths and summaries instead.
From PDF to Ground Truth Dataset: Extraction Traps That Break Evidence Anchors
In a source-grounded benchmark, extracted text is the evidence. How pdftotext modes, OCR damage and pseudonymisation quietly break evidence anchors.
Measuring the Resolution of a 50-Question Benchmark
Benchmark statistical significance in practice: confidence intervals, minimum detectable gaps, and a fairness check on a 50-question AI benchmark.
An LLM Judge That Can't Invent a Score
AI model evaluation with an LLM judge: the judge returns a JSON block the harness checks against the rubric, and an unscored answer beats a guessed one.
Position Bias in Pairwise LLM Evaluation
Asking the judge twice with answers swapped exposes position bias. The scoring table, worked examples, and why pairwise verdicts can't build a leaderboard.
Testing the Rubric Before You Test the Model
Calibration fixtures test a rubric before it grades any model. The dominant defect is double credit across the gate: here's what our fixture runs found.
Generating a 50-Question Benchmark With One Approval Step
A one-prompt authoring flow that stops once at the blueprint, runs four machine gates, and resumes from its own files. What it catches, and what it doesn't.