Top Banner
#MozCon Justin Briggs • Getty Images Talking Back to Conversational Search @justinrbriggs • [email protected]
50

Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Apr 16, 2017

Download

Marketing

Justin Briggs
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#MozCon

Justin Briggs • Getty Images

Talking Back to Conversational Search

@justinrbriggs • [email protected]

Page 2: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Working in SEO

Page 3: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 4: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 5: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 6: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 7: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 8: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Exciting time for SEO

Page 9: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Web Search

Page Page Page

Search in web, against documents

Page 10: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 11: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

“What is limburger”

Page 12: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

“Play say it ain’t so”

Page 13: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

“Schedule a meeting with Lynn”

Page 14: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

“Navigate to grandma’s house”

Page 15: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

These queries never touch the web

Page 16: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Search

Web Email

Search against capabilities

Knowledge

GraphSchema

Mobile

Apps

Documents Databases Applications

Page 17: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 18: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Understanding natural language

Breaking down strings #1 Tokenization #2 Parts of speech tagging #3 Lemmatization

#4 Named entity detection

Page 19: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#1 Tokenization

Who directed pulp fiction

Page 20: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#2 Parts of speech tagging

Page 21: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#2 Parts of speech tagging

Who directed pulp fiction

WP VBD NNP

WP: wh-pronoun

VBD: verb, past tense

NNP: noun, proper, singular

Page 22: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#3 Lemmatization

am, are, is => be

car, cars, car’s, cars’ => car

Page 23: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#3 Lemmatization

Converting to canonical words

Page 24: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#4 Named entity detection

photos of Emma Watson

Page 25: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

opennlp.apache.org

String tokens[] =

tokenizer.tokenize("An input sample sentence.");

"An", "input", "sample", "sentence", "."

Page 26: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

opennlp.apache.org

Photos of Emma Watson

Photos of <START:person>Emma Watson<END>

Page 27: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Knowledge graph capability

Page 28: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Replicate using mql

https://www.googleapis.com/freebase/v1/

mqlread?query={<insert query here>}

Page 29: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Strings to things

[{

"!/film/film/directed_by": [{

"/type/object/name": "pulp fiction",

"/type/object/type": "/film/film"

}],

"/type/object/name": [{}],

"/type/object/type": "/people/person"

}]

Question: Who directed the movie Pulp Fiction?

Page 30: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Answering queries without pages

{

"result": [

{

"!/film/film/directed_by": [

{

"/type/object/type": "/film/film",

"/type/object/name": "Pulp Fiction"

}

],

"/type/object/type": "/people/person",

"/type/object/name": [

{

"lang": "/lang/en",

"type": "/type/text",

"value": "Quentin Tarantino"

}

Answer: Quentin Tarantino

Page 31: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

quepy.machinalis.com

Page 32: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 33: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Curating entity content

Page 34: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Attributes and connections

Emma Watson

Born: April 15, 1990 (age 24)

From: Paris, France

Page 35: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Which Brad Pitt?

Page 36: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Next level keyword targeting

Actor: actor, producer, Angelina Jolie, 1963, Fight Club

Boxer: martial arts, boxing, Olympics, 1981

Page 37: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Find common attributes by type

Page 38: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Dynamic targeting using entity attributes

<Name> (# <Number>) : <Position> for <Team>in <City>, <State>

Page 39: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Dynamic targeting using categories

Emma Watson Photos - News, Publicity, Event, & Premiere Pictures

!Seattle Seahawks Photos - Team, Player, & Fan Pictures

!Justin Bieber Photos - News, Concert, & Fan Pictures

!

Page 40: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 41: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Marking up schema

Page 42: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Schema takes you from web to data

Page 43: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

JSON-LD: JSON-based linked data format

<script type="application/ld+json">

{

"@context": "http://schema.org",

"@type": "Person",

"name": "John Doe",

"jobTitle": "Graduate research assistant",

"affiliation": "University of Dreams",

"url": "http://www.example.com",

"address": {

"@type": "PostalAddress",

"streetAddress": "1234 Peach Drive",

"addressLocality": "Wonderland",

"addressRegion": "Georgia"

}

}

</script>

Page 44: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 45: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014
Page 46: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Exciting time for SEO

Page 47: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Feels like we’re over here

Page 48: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

We’re actually over here

Page 49: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

Getting to AI faster

Page 50: Conversational Search, Entities, and Knowledge Graph - Mozcon 2014

#MozCon

Justin Briggs • Getty Images@justinrbriggs • [email protected]