AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Importance of Auto Data Validation in Machine Learning
I think that's really only useful if you care about monitoring your pipelines. If some human is validating all of the outputs, for example, like if a CEO is looking at the dashboard, that's very different use case than fraud detection or like something that is not manually looked at for every output. Then this becomes relevant. And I guess my other question is- Well, there's a lot of real production models that aren't just that kind. Like a real production model also generates insights to present to the CEO every, I don't know, like month or so. That's fine. But if they're getting requests on a regular basis and it's making predictions, you're ret