Defining Diversity

7
Defining Diversity Marko Grobelnik Jozef Stefan Institute

description

Tutorial at the RENDER Kick-off Meeting, Jozef Stefan Institute

Transcript of Defining Diversity

Page 1: Defining Diversity

Defining Diversity

Marko GrobelnikJozef Stefan Institute

Page 2: Defining Diversity

Wikipedia on Diversity

Page 3: Defining Diversity

Dictionary definition

Page 4: Defining Diversity

What could be diversity in RENDER?

o RENDER is a lot about • Communication between people, • …documents of different kinds,• …their usage, and• …in particular, their perception• …in particular, their perception

o Informally speaking, RENDER is in many ways dealing with modeling of “soft” issues within (mostly textual) communication

Page 5: Defining Diversity

Attempt to structure the notion of “diversity” within RENDER

o Who are the main “actors” within RENDER scenarios?• Publishers producing content

� Main stream news, blogs, forums, twitter, …

• Users producing usage data• Users producing usage data� Passive readers, Prosumers (consumer/provider)

Page 6: Defining Diversity

Diversities coming out of “content”

o Topic• What topic is text about?

(categorization, segmentation)o Social

• Who is writing? (publisher, author)• Who is being written about? (people,

organizations)• Influence

o Geographical

o Knowledge• Fact coverage (difference in

sources)• Relationship extraction (how

entities are related)o Cross-lingual / Multi-lingual

• What language is being used?• Cross-lingual story linkingo Geographical

• Where the content was produced?• Which geography is being

addressed?o Opinion

• Sentiment (positive/negative/neutral)• Opinions (beyond polarized sides)• Reporting bias (differences)

• Cross-lingual story linkingo Context

• Temporal context (story linking, discussion threads, trends in other dimensions)

• Background knowledge (knowledge bases, ontologies, …)

• Contrasting with other sources (e.g. main-stream vs. twitter)

Page 7: Defining Diversity

Diversities coming out of “usage”

o Demographic context• Who is the user? (age, gender, job, income)

o Topic• What are user interests? (predefined, calculated)

o Geography• Where a user is coming from? (home, accessing)

o Access method • How a user is accessing data? (web browser, mobile, forum, phone, email)• How a user is accessing data? (web browser, mobile, forum, phone, email)

o Social context• With whom a user is connected to? (social network, communication)

o Time• When a user is accessing information (absolute, day of week, hour of day)

o Historical• What a user was doing in the past?• How user activities change trough time? (trends)