A dataset from the Geosteering World Cup 2021 containing 176 expert readings of the same known geological scenario. The data likely captures variations in expert interpretation and labeling. It was created for a 2021 competition focused on geosteering.
Use Cases
- Study inter-annotator agreement and label noise based on multiple expert readings.
- Benchmark label aggregation or noise-cleaning algorithms based on known ground truth.
- Analyze expert decision-making patterns in geosteering based on repeated interpretations.
- Train models to predict annotation uncertainty based on expert disagreement.
Strengths
- Known ground truth for the underlying geology, enabling direct noise measurement.
- Involves 176 expert annotators, providing a substantial sample of expert opinion.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Last update date is unknown; freshness unverified.
Provenance
- Source
- Geosteering World Cup 2021
- Collection Method
- Likely contains expert annotations collected for a competition.
- Time Range
- 2021