On: How to fit any dataset with a single parameter

[bflesch]

This sounds like it could explain many ML models ;)

[mr_toad → bflesch]

It’s actually possible in theory for part of a neural net to simply encode the training data and use it to regurgitate responses. In practice, more efficient parts of the network should win in the end.

[NavinF → mr_toad]

Has nobody in this thread heard of https://en.m.wikipedia.org/wiki/Cross-validation_(statistics...

Every ml tutorial start with splitting your data in into a training set and validation set.