This is nice, thanks. I will probably point to this as a brief overview.
I would argue that because of AIXI’s treatment of the environment program as two-part code (alternatively, its formulation using a Bayesian mixture of stochastic environments) it actually does “believe in objective randomness” insofar as that is a meaningful concept.
This is nice, thanks. I will probably point to this as a brief overview.
I would argue that because of AIXI’s treatment of the environment program as two-part code (alternatively, its formulation using a Bayesian mixture of stochastic environments) it actually does “believe in objective randomness” insofar as that is a meaningful concept.