Careful formulations of Occam’s Razor have always spoken of bits of parameters, rather than count of parameters, for basically this reason: a “parameter” could be a boolean yes/no, or a number, or a five-node digraph, or… pretty much anything you can write down. This paper demonstrates that point very well, but I don’t think the point itself is new.
Careful formulations of Occam’s Razor have always spoken of bits of parameters, rather than count of parameters, for basically this reason: a “parameter” could be a boolean yes/no, or a number, or a five-node digraph, or… pretty much anything you can write down. This paper demonstrates that point very well, but I don’t think the point itself is new.
Well put.