Automatic Construction of a semantic graph The Sunflower Algorithm - - PowerPoint PPT Presentation

automatic construction of a semantic graph
SMART_READER_LITE
LIVE PREVIEW

Automatic Construction of a semantic graph The Sunflower Algorithm - - PowerPoint PPT Presentation

Automatic Construction of a semantic graph O. Chabrol, D. Norrestam Project background Entity Disambiguation Automatic Construction of a semantic graph The Sunflower Algorithm Q-numbers EDAN70 Project in Computer Science Implementation


slide-1
SLIDE 1

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Automatic Construction of a semantic graph

EDAN70 Project in Computer Science

  • O. Chabrol
  • D. Norrestam

May,23rd 2016

slide-2
SLIDE 2

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-3
SLIDE 3

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-4
SLIDE 4

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Entity Disambiguation

◮ Basic problem : different entities can have the same

name

slide-5
SLIDE 5

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-6
SLIDE 6

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Goal of the algorithm

◮ Find concepts linked to every entity ◮ Using categories linked to article in database

slide-7
SLIDE 7

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

◮ Find the ”best” categories ◮ We take the most recurrent categories as best categories

slide-8
SLIDE 8

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-9
SLIDE 9

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Q-numbers

◮ Allows the link between languages

slide-10
SLIDE 10

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Q-numbers

◮ Unique identifier for entities ◮ Every entity (articles and categories) has one universal

Q-number

slide-11
SLIDE 11

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Q-numbers examples

slide-12
SLIDE 12

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-13
SLIDE 13

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Overview

The algorithm consists of the following steps:

  • 1. Extract information from Wikipedia
  • 2. For each language

2.1 Parse information 2.2 Create data structures

  • 3. Merge languages and create semantic graph
slide-14
SLIDE 14

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-15
SLIDE 15

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Extract information

◮ Have two options

  • 1. Parse Wikipedia
  • 2. Use existing dumps

◮ Chose existing dumps

◮ DBpedia (2015-10)

slide-16
SLIDE 16

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-17
SLIDE 17

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Parsing the dumps

◮ for every language we have : ◮ and we create :

slide-18
SLIDE 18

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-19
SLIDE 19

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Merging data

◮ after merging the three languages: ◮ we can select the five ”best” categories ◮ with 123 languages

slide-20
SLIDE 20

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Outline

Project background Entity Disambiguation The Sunflower Algorithm Q-numbers Implementation Overview Extract information Parsing the dumps Merging data Depth Demonstration

slide-21
SLIDE 21

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Depth

◮ we already have a width concept ◮ categories are also part of more general categories ◮ which leads to depth

slide-22
SLIDE 22

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Computation of ratios

slide-23
SLIDE 23

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

Demonstration

Interface example

slide-24
SLIDE 24

Automatic Construction of a semantic graph

  • O. Chabrol,
  • D. Norrestam

Project background

Entity Disambiguation The Sunflower Algorithm Q-numbers

Implementation

Overview Extract information Parsing the dumps Merging data Depth

Demonstration

End

End of the presentation Thank you ! Any questions ?