Do science fair judges care about real world-feasability?

So i made a ML model to diagnose a disease for way cheaper, way more accessible, and more accurate than it is currently. Therefore, I can make big claims on how impactful it could be. I can prove these claims with the performance on the validation set and by just estimating the cost. However, I don't actually have a way of like applying this in the real world, so can i still make big claims like this when i don't have real world proof of how this would work?

Also, i'm 90% sure my model isn't overfit i did everything i could to minimize that including strong reg, validation set, smote, etc, and stuff so the accuracy is legit

3 comments