# Using a teammate's geological interpretation as training labels

**URL:** https://community.drivendata.org/t/using-a-teammates-geological-interpretation-as-training-labels/11543
**Category:** GEMS Prize Challenge
**Created:** [September 25, 2026, 4:43pm UTC](https://community.drivendata.org/t/using-a-teammates-geological-interpretation-as-training-labels/11543 "2026-09-25T16:43:10Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![exposed](https://avatars.discourse-cdn.com/v4/letter/e/e0b2c6/32.png) [@exposed](https://community.drivendata.org/u/exposed)
#### Post date: [September 25, 2026, 4:43pm UTC](https://community.drivendata.org/t/using-a-teammates-geological-interpretation-as-training-labels/11543/1 "2026-09-25T16:43:10Z")

</div>

Hi, the rules encourage teams that combine geoscience and data science. If a geologist on our team reviews our model’s candidate faults on the public lidar (accept or reject each one, and occasionally trace a fault), may we use those judgments as training labels or hard negatives for the model, provided the final submission is produced entirely by our code? Is there any restriction on doing this for areas inside the GeoDAWN region? Thanks!