Gabriel Poesia

‘Ideal learning’ of natural language: Positive results about learning from positive evidence (@ Journal of Mathematical Psychology 2007)

Nick Chater, Paul Vitányi

Link

Paper with a formal argument showing the possiblity of learning arbitrary languages in the limit from just positive examples when (1) the learner's hypothesis space includes all possible computations, and (2) the examples are drawn from a fixed distribution (i.e., not adversarially generated).