My feed today showed this paper published by Buehrer and colleagues (2017) with respect to a mathematical framework for superintelligent machines.
Leaving the sophisticated title aside, their paper is about a class calculus expressive enough for self-learning optimization (it is capable of describing and improving its learning process). As specified in the paper:
"It can design and debug programs that satisfy given input/output constraints, base d on its ontology of previously learned programs. It can improve its own model of the world by checking the actual results of the actions of its robotic activators." [source]
Some potential applications and examples to which this abstract concept can be used for are (as they suggest):
- check the black-box of a crashed car to determine the reason of the crash:
"...if it was probably caused by electric failure, a stuck electronic gate, dark ice, or some other condition that it must add to its ontology in order to meet its sub-goal of preventing such crashes in the future." [source]
Now, to nudge you and to provide some sort of interest for reading the paper and maybe understanding some of its implications is that the paper is a description of something called 'The Master Algorithm' - by machine learning researcher Pedro Dominguez. Now you're interested?
To stay in touch with me, follow @cristi
Cristi Vlad Self-Experimenter and Author