Boost Your AI With Quality Data
All AI projects start with data. No matter how simple your idea is, you cannot develop machine learning algorithms without examples to train them on. And after the first prototype, you start chasing better metrics and find that the amount and quality of your data is crucial. That is when a good data labeling pipeline comes into its own.
In this talk we introduce data labeling pipelines. We’ll show you how they’re used right now in areas like search relevance, content moderation, voice assistants and self-driving cars. We will explain how to fight concept drift in machine learning, how to build complex products using the human-in-the-loop model, and how not to be driven mad by people management. Beginners and experienced specialists alike can learn something from this talk.