Evaluation may be easier than generation (extended abstract (1996) [1 citations — 0 self]
Abstract:
Kearns et al. [18] defined two notions for learning a distribution D. The first is with generator, where the learner presents a generator that outputs a distribution identical or close to D. The other is with an evaluator, where the learner presents a procedure that on input x evaluates correctly (or approximates) the probability that x is generated by D. They showed an example where efficient learning by a generator is possible, but learning by an evaluator is computationally infeasible. Though it may seem that generation is, in general, easier than evaluation, in this paper we show that the converse may be true: we provide a class of distributions where efficient learning with an evaluator is possible, but coming up with a generator that approximates the given distribution is infeasible. We also show that some distributions may be learned (with either a generator or an evaluator) to within any ffl? 0, but the learned hypothesis must be of size proportional to 1=ffl. This is in contrast to the distribution-free PAC model where the size of the hypothesis can always be proportional to log 1=ffl.
Citations
| 89 | When won't membership queries help – Angluin, Kharitonov - 1995 |
| 30 | How to Generate Factored Random Numbers – Bach - 1988 |
| 29 | Goldwasser: New paradigms for digital signatures and message authentication based on non-interactive zero-knowledge proofs – Bellare, Shafi - 1990 |
