Limits of Training
Understanding the boundaries of AI training
If all examples are the same, the model does not learn anything new. Just like a student does not learn more by solving the same task again and again.
If examples are wrong, the model can learn the wrong thing. Even if the math improves, the real-world result will be incorrect.
If the model is too complex for the amount of data, it cannot learn properly.
If the model is too simple, it cannot learn important edge cases.