On the limit of some algorithmic approach to circuit lower bounds

Abstract

We propose a framework — generic algorithm framework for studying the limit of a certain algorithmic approach for showing circuit lower bounds. We show some illustrative examples for explaining the  motivation/justification of our framework while leaving one key technical and very challenging problem open. We hope that this framework is useful for investigating the limit of algorithmic approaches to define computationally hard problems.