Session Outline
Codon developed a content moderation system for a Swedish social media app. In this talk, we’ll explain how we tackled challenges like limited and biased data, changes in trends (concept drift), and managing expectations. The model, which has been running for two years, is still active and regularly updated. We’ll also share how we keep it up to date.
Key Takeaways
- Data augmentation for textual data
- Metrics and expectation management
- Prediction confidence estimation
