| Abstract |
In the LarKC project, we are building the Large Knowledge Collider, an engine for infinitely scalable reasoning with very large data-set (think: web-scale). We plan to achieve this infinite scalabilty through parallelisation/distribution and through anytime/approximate reasoning.
Both for distributed reasoning and for anytime reasoning, it is important to know when to stop (in particular when your dataset is too large to be handled completely anyway).
Cognitive scientists are now helping us to find out which stopping rules people (and other animals) are using when they have to decide when to stop a task (or when to switch from one task to another). The hope is that they will uncover general stopping-heuristics that are not only effective for animals, but that can also be used for machine computations.
In this talk I will outline the general ideas of this research, describe what is already known about stopping rules, and what there still is to be found out.
|