Has anone used a Probably Approximately Correct (PAC) framework to develop a rough formalization of taste of a learner A(S)? [and whether a learner “has it in them” to correctly classify H, a complex hypothesis space that is more complicated than the S they’ve seen?]
Especially learners who come from “shitty environments” (have poor S), but still have the A(S) in them to suddenly “jump” once they have exposure to the right people
Some “A” functions might have unusually nonlinear behavior once a person is exposed to the right set of tutors or environment (and some “A” functions never have it in them)
Has anone used a Probably Approximately Correct (PAC) framework to develop a rough formalization of taste of a learner A(S)? [and whether a learner “has it in them” to correctly classify H, a complex hypothesis space that is more complicated than the S they’ve seen?]
Especially learners who come from “shitty environments” (have poor S), but still have the A(S) in them to suddenly “jump” once they have exposure to the right people
Some “A” functions might have unusually nonlinear behavior once a person is exposed to the right set of tutors or environment (and some “A” functions never have it in them)