Surge AI Blog company blog
30% of Google's Emotions Dataset Is Mislabeled
Last year, Google released their “GoEmotions” dataset: a human-labeled dataset of 58K Reddit comments categorized according to 27 emotions. The problem? A whopping 30% of the dataset is mislabeled! Check out some of the egregious errors, and learn how to build better datasets.30% of Google's Emotions Dataset is…
- Points
- 334
- HN comments
- 144
- Submissions
- 1
- Cited by
- 0 commenters
AI & Machine LearningDatabases & Data2022
On Hacker News
| Submitted | Title on HN | Points | Comments |
|---|---|---|---|
| Jul 14, 2022 | 30% of Google's Emotions Dataset Is Mislabeled by echen | 334 | 144 |
