4 ms·
You want to use a razor to shave a bear. Simplicity is not enough to favor a belief model. Machine Learning has shown us that large complex models can perform m
by goldenkey 5y ago
You want to use a razor to shave a bear. Simplicity is not enough to favor a belief model. Machine Learning has shown us that large complex models can perform much better at providing the ground truths.
https://en.wikipedia.org/wiki/Philosophical_razor https://en.wikipedia.org/wiki/Philosophical_razor
Razor's aren't scientifically rigorous theorems. Feel free to use them but don't argue by them in a public forum, it's disingenuous to truth.
- morpheos137 5y agoPretty good point. People resort to occams razor as if it is a fundamental constant of nature valid in all cases. Occams razor has heuristic value not probitive value.