5 ms·
> The real function of Occam's Razor is to allow us to simplify models for our own benefit (because simpler models are easier to interpret and use), not necessa
by Bromskloss 9y ago
> The real function of Occam's Razor is to allow us to simplify models for our own benefit (because simpler models are easier to interpret and use), not necessarily because the simple answer is the Actual Truth(TM), which we can't know outside of observation.
What about assigning greater prior probability to hypotheses that are simpler in some sense? Is there perhaps something to that too?
- lmm 9y agoI'd say it's legitimate, just because simpler models have turned out to be true more often in the past, as best as we can tell.
- Bromskloss 9y agoMaybe even _that_ reasoning relies upon simple hypotheses being more likely. In this case, the simple, and therefore favoured, hypothesis would be that "models with a simplicity that is similar to that of other successful models are more likely to be true".
- lmm 9y agoThat sounds like a more complex hypothesis to me.
- Bromskloss 9y agoMaybe it's simpler than "models with a simplicity that is dissimilar to that of other successful models are more likely to be true", though it's not obvious what should be considered simple and not.
- lmm 9y agoKolgomorov complexity seems like the obviously correct measure, though AIUI it's only defined relative to some fixed Turing machine. I have an idea of doing a PageRank-style weighting of Turing machines, based on how simple/complex simulating one under another is. That would align with my intuition that most basic universal Turing machines are equally valid, but one that has (say) a short instruction that outputs one particular long sequence of symbols is probably not a valid machine to use.
- balabaster 9y agoAssuming that other successful models are also true and not just a wolf in sheep's clothing.
- Bromskloss 9y agoLet's forget that I said "true", and switch it to "successful" instead.