What does useful mean? I would argue most models are useless without a good understanding of the parameters going into that model. That is where the domainspeci

What does useful mean? I would argue most models are useless without a good understanding of the parameters going into that model. That is where the domainspecific knowledge comes in. Fitting some curve to data is useless unless you know what the parameters correspond to, and how they influence the parameters of the curve.