User talk:Aba700

From Master Projects
Jump to: navigation, search


has title::Large scale reasoning using models of human cognition
status: ongoing
Master: project within::Cognitive Science
Student name: student name::Arjon Buikstra
number: student number::1536605
Dates
Start start date:=2010/08/01
End end date:=2010/10/17
Supervision
Supervisor: Frank van Harmelen and Annette ten Teije
Second reader: has second reader::Lael Schooler and Hansjoerg Neth
Company: has company::VU & ABC at Max Plank Institute for Human Development in Berlin
Poster: has poster::Media:Media:Posternaam.pdf

Signature supervisor



..................................

Abstract

Introduction

This project aims to benefit both the reasoning over large scale semantic web repositories, and to investigate human behavior through (existing) cognitive models using real-world memory sets. To reach this goal, two known frameworks are further integrated (see also Martijn Brakenhoff's work), namely LarKC and ACT-R. LarKC is a system that was created in the past few years in Europe for querying large data repositories with Semantic web techniques (the data is handled inside the program as rdf triples, queries are written in SPARQL). The source is freely available, and work is still being done (also at the VU) to improve the system, and to create new plug-ins. ACT-R is a cognitive modeling framework, originally devised by Anderson at Carnegie Mellon around the 80's. It's still being used today by cognitive scientists wishing to model human tasks (e.g. simple recall tasks with motor modeling), but with the downside that its memory is often fairly limited and specifically made for the task. Martijn Brakenhoff already made an integrated version of both systems in Java during his MSc thesis, so a big part of this project will be to explore the improved functionality of this hybrid for both areas of research.