Data collection is a very error-prone process. And the important fact is even a little noise in the data can cause chaos in your model. Our recent publication, "Deep Cleaner - A Few Shot Image Dataset Cleaner using Supervised Contrastive Learning" deals with the same.
(1/n)
In India, we usually don’t take precautions until something serious happens. Both people and the government tend to ignore problems, even when they know these could lead to major consequences in the future. I really wish this mindset changes someday 🤞