E-MaLeS 1.1. Picking the right search strategy is important for the success of automatic theorem provers. E-MaLeS is a meta-system that uses machine learning and strategy scheduling to optimize the performance of the first-order theorem prover E. E-MaLeS applies a kernel-based learning method to predict the run-time of a strategy on a given problem and dynamically constructs a schedule of multiple promising strategies that are tried in sequence on the problem. This approach has significantly improved the performance of E 1.6, resulting in the second place of E-MaLeS 1.1 in the FOF divisions of CASC-J6 and CASC$@$Turing.

Keywords for this software

Anything in here will be replaced on browsers that support the canvas element