Claimsweb

The “Claims Web”: a starting point for Social Semantic Argumentation

Jodi SchneiderWork in Progress presentation

DERI Social Software Unit2011-08-05

Second Year GRC Report2011-07-15Galway, Ireland

2

Social Semantic Web

Argumentation

The “Claims Web”: a starting point for Social Semantic Argumentation

A tale of two whiskeys

Jodi SchneiderWork in Progress presentation

DERI Social Software Unit2011-08-05

Second Year GRC Report2011-07-15Galway, Ireland

What’s the best whiskey to buy, on a budget?

• What’s a good Irish whiskey to bring as a hospitality gift?

• Do you have any advice for alcohol that's mainly available in the U.S.? Not super-expensive, but kind of nice?

Which answer would you rather get?

• I had a little think about it, and I came up with a list of options for you. ….

OR• http://ask.metafilter.com/131843/What-is-you

r-best-favorite-bourbon-for-under-100

Need for “sensemaking”There are a growing number of conversations on

the Social Web, but we do not take advantage of all the social knowledge and social belief.– There are no readily-available summaries for

threads of conversation, meaning we must wade through redundant posts, often with poor rhetoric (e.g. circular reasoning).

– Multiple disparate conversations may occur on the same topic, and there is no easy way to bring these together.

– Hard to identify people with similar, different opinions from me

9

Building a Web of CLAIMS

• What if instead of following mailing lists, blogs, online magazines, scientific journals…

• You could search for and follow CLAIMS?– What are the important claims about “Irish whiskey”? (or “climate

change” or “sensor networks” or…)?– What are the points of disagreement?– Who is making these claims? – Should I believe them?

• Do they have conflicts of interest?• Are they trustworthy? Do people I trust, trust them?• Are their ideas credible?• Do they have similar needs? Do the same aspects and features of the

problem matter to them?

Why build the Claims Web?

• What do you do when people disagree?• Getting meaning from inconsistent knowledge

– Relevant for collaborative filtering, decision-making, sensemaking, …

• Claims & Reasons are fundamental… and need to be connected!

10

Moving towards a Web of Arguments

• World Wide Argument Web (WWAW) isn’t here yet– “a large-scale Web of interconnected arguments posted by

individuals to express their opinions in a structured manner'’ (Rahwan 2007).

– RDFS and OWL, Argument Interchange Format (AIF).• Problems with the WWAW for the Social Web

– AIF requires argument schemes – overly complex!– Users don’t explicitly mark their positions. Need to infer.

• Claims Web will move us towards the WWAW– Extract info from users’ posts.– Cover only the Social Web.– Focus on claims & reasons.

11

Existing argumentation technology is complex

• Argument schemes (e.g. Walton’s) for analysts• Argument Interchange Format

GOAL: Bridge existing work to suit *humans*, informal discussions

14

Approaches

• Mixed-initiative systems– Natural language processing to discover/infer

claims & reasons– Executable English to make it easy to specify– Games with a purpose to encourage people

Process

Detailed Process

Build a web of claims to:• Gather knowledge from the

inconsistent opinions of a crowd• Integrate argumentative conversations

across the silos of the social web• Provide support for socially-embedded

sensemaking and decision-making

Questions? Comments?

19

Example argument from Wikipedia deletion discussions