Every year, fraud costs the global economy over $5 trillion. AI practitioners are on the front lines of this battle building and deploying sophisticated ML models to detect fraud, saving organizations billions of dollars in the process. Of course, it’s a challenging task as fraud takes many forms and attacks vectors across industries. ML teams need an approach that is both reactive in monitoring key metrics and proactive in measuring drift, counter-abuse ML teams.
In this webinar, Reah Miyara, Arize’s Head of Product and former Google AI lead for algorithms and optimization, will cover best practices in ML observability for fraud models.
In this webinar, you’ll learn best practices for how to:
- Account for model, feature and actuals drift to ensure your models stay relevant
- Troubleshoot performance degradations across various cohorts
- Avoid common pitfalls from misleading evaluation metrics to imbalanced datasets