How to Build a Scalable System for Data Annotation | Expert Tips
Nothing is loaded from YouTube until you press play, and then without tracking cookies.
About this video
Discover how to build a scalable system for data annotation to handle malicious actors and subjective work effectively. Learn the key strategies and technologies behind successful data annotation projects. #DataAnnotation #ScalableSystem #DataAnnotationTips #ExpertAdvice #Technology #MachineLearning #DataScience #DataAnnotationProjects #DataManagement #SubjectiveWork #NEARprotocol #NEARai
Transcript
159 wordsCaptions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.
- 0:00
ideally if you if you want toate you know millions of examples of data uh and it requires people doing it you have to you have to like like build a scalable system where there are going to be a lot of malicious actors uh without an ability it's almost like designing a blockchain with a differ on blockchain at least certain things are verifiable like you know like I I can say that this message is signed by that person while
- 0:23
in annotating data everything is subjective right so you have thousands of people doing subjective work verifying subjective work and obviously like for a particular you know like there's a uh yeah there have there have been some large scale efforts and uh like you will never be able to document every single possible interaction right so people need to convert to some uh to some level
- 0:46
of uh uh agreement between themselves right so it's a