Hyperparameters play a significant role in bias when you're dealing with imbalanced classes, or long tail samples. But this ties back to the original data prob

Hyperparameters play a significant role in bias when you're dealing with imbalanced classes, or long tail samples. But this ties back to the original data problem, right? If you don't have enough training samples for (known or unknown) unknowns, your model is likely to be biased against them.