Enabling Cognitive Mechanisms in Web-scale Reasoning
|has title::Enabling Cognitive Mechanisms in Web-scale Reasoning|
|Master:||project within::Knowledge Technology and Intelligent Internet Applications|
|Student name:||student name::Martijn Brakenhoff|
|Supervisor:||Annette ten Teije|
|Second reader:||has second reader::Frank van Harmelen|
|Company:||has company::Center for Adaptive Behavior and Cognition - Max Plank Institute for Human Development - Berlin|
The Semantic Web aims to make the World Wide Web machine readable, but reasoning isn't very scaleble. The LarKC platform hopes to enable reasoning at web-scale. The platform has to deal with large and messy data-sets. We humans have to do the same every day. In this thesis we will combine the LarKC platform with a cognitive architecture called ACT-R. We will perform a high-level analysis of the two architectures, attempt to find a mapping from one architecture to the other and implement it.
The Semantic Web aims to make the World Wide Web machine readable, but reasoning doesn't scale easily. The Large Knowledge Collider Project (LarKC) aims to provide a platform for reasoning at web-scale. The platform has to deal with multiple, large, messy and inconsistent data-sets. The human mind faces similar problems every day.
In this thesis we integrate the LarKC platform with a cognitive architecture in order to examine aspects of human reasoning that might be beneficial when reasoning at web-scale. We performed a high-level analysis of the two architectures, briefly explored and experimented with human stopping rules. We propose and implement a prototype integration, create a use-case and add stopping rules based on our experiments.