Evolutionary & Swarm Computing for the Semantic Web

Post on 19-Jan-2015

92 views 6 download

Tags:

description

Semantic Web will be the next big thing in the world of internet. This presentation talks about various approaches that can be used to query the underlying triple store that has all the information.

transcript

Evolutionary & Swarm Computing for the Semantic Web

By:

ANKIT A SOLANKIANUJ IYER

PRATIK K SHAH

What is Web x.0

• Web 1.0• Web 2.0• Web 3.0 also known as semantic web.

Semantic Web

Why we need it?Semantic Search. I say ‘google’ is dumb!

Yes I typed there ‘google’

Facebook as an example.

Concepts

• Triples- Subject Predicate Object.• Ontology• Swarm Computing• Evolutionary Computing• Local Search

A few important challenges

• Storage• Time• Storage vs Time• Distributed nature• Ownership

Evolutionary Algorithms for Querying

http://www.seoskeptic.com/wp-content/uploads/2011/01/rdf-triple.jpg

a)Evolutionary Algorithms for Querying

1) ASK2) GET3)Optimizer

http://www.seoskeptic.com/wp-content/uploads/2011/01/rdf-triple.jpg

eRDF Framework

ASK

GET(<*,p,o>)GET(<s,*,o>)GET(<s,p,*>)

GET

• GET(<*,p,o>)• GET(<s,*,o>)• GET(<s,p,*>) <s, p, o>

A simplified bookstore data (dataset“A”)

ID Author Title Publisher YearISBN0-00-651409-X The Glass Palace 2000id_xyz id_qpr

ID Name Home Page

ID CityHarper Collins London

id_xyz Ghosh, Amitav http://www.amitavghosh.com

Publ. Nameid_qpr

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

1st: export your data as a set of relations

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Another bookstore data (dataset “F”)A B D E

1 ID Titre Original

2

ISBN0 2020386682 A13 ISBN-0-00-651409-X

3

6 ID Auteur7 ISBN-0-00-651409-X A12

11

12

13

TraducteurLe Palais des miroirs

NomGhosh, AmitavBesse, Christianne

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

2nd: export your second set of data

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

3rd: start merging your data

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

3rd: start merging your data (cont.)

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

3rd: merge identical resources

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Relations can be generated on-the-fly at query time

via SQL “bridges” scraping HTML pages extracting data from

Excel sheets etc.

Conflict in f:author V/S a:author

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Merge with different datasets

e.g., the “dbpedia” project can extract the “infobox” information from Wikipedia already…

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Merge with Wikipedia data

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Merge with Wikipedia data

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Simple SPARQL example

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Simple SPARQL exampleSELECT ?isbn ?price ?currency # note: not ?x!WHERE { ?isbn a:price ?x. ?x rdf:value ?price. ?x p:currency ?currency. FILTER(?currency == € }

SELECT ?isbn ?price ?currency # note: not ?x!WHERE { ?isbn a:price ?x. ?x rdf:value ?price. ?x p:currency ?currency. FILTER(?currency == € }

Returns: [[<..409X>,50,€], [<..6682>,60,€]]

Reference: http://www.w3.org/2009/Talks/1030-Philadelphia-IH/

Optimizer

Evolutionary Algorithm

a)Population Evaluation

b)Survivor Selection

c)Offspring Generation

b)Computing for Logical Entailment

• Swarm computing – Ants, Bees, Termites etc

Ant Colony Optimization

Reasoning as a graph traversal

Reasoning as a graph traversal

Reasoning Agent

• Agent can have schema triple in memory to do reasoning• Movement routing can be done either

– Based on properties– Based on pheromones– Random

• To control movement, use happiness function

Proof of concept

THANK YOU