Yet Another Parallel Hypothesis Search for Inverse Entailment.
Hiroyuki Nishiyama, Hayato Ohwada · 2015
In this study, we design and implement a powerful Inductive Logic Programming (ILP) system to conduct a parallel hypothesis search for inverse entailment. One of the most important parts of ILP is speeding up the hypothesis search, and a number of parallel hypothesis exploration methods have been proposed. Recently, the Map-Reduce algorithm has been used for large-scale distributed computing, but it is difficult to apply such cloud technology to the ILP hypothesis search problem. By contrast, we designed a method that can dynamically distribute tasks for the hypothesis search, and implemented a network system in which modules autonomously cooperate with each other. We also conducted a parallel experiment on a large number of CPUs. Results confirm that the hypothesis search time is shortened according to the number of computers used, without reducing the optimality of the generated hypothesis.