Photo by Dung Anh on Pexels

A recent Reddit thread sparked a conversation among AI researchers regarding data collection and preparation strategies for machine learning (ML) and reinforcement learning (RL) projects. The discussion, initiated by a user curious about common practices, touched upon methods like self-labeling, leveraging open datasets, and outsourcing annotation. Participants explored the various challenges and time investments associated with creating datasets for both academic and independent research. The full Reddit discussion can be found at: https://old.reddit.com/r/artificial/comments/1o5salp/how_do_you_usually_collect_or_prepare_your/