#epistemology #ai # [[Epistemic status]] #shower-thought # Criterion of demarcation ![[DALL·E 2022-06-23 09.08.21 - A man tweaking the robot's brain, digital art.png]] How does [[The criterion of demarcation]] applies to [[Computing/Intelligence/Machine Learning/Machine Learning|ML]]? It is called optimisation, what updates your beliefs towards the [[Signal]]s while moving away from [[Philosophy/Rationality/Anti model/Noise|noise]]. ${\displaystyle H(p,q)=-\operatorname {E} _{p}[\log q]}$ For example, we can use cross-entropy[^1] as a loss function[^2] to update our belief towards reality, in the case of machines, to measure the model performance. ![[DALL·E 2022-06-23 09.13.06 - A man separating a blue liquid from a red one within the robot's brain, digital art.png]] # External links [^1]: https://en.wikipedia.org/wiki/Cross_entropy [^2]:https://ml-cheatsheet.readthedocs.io/en/latest/loss_functions.html