10
Web 3.0 Combining Domain Intelligence with Social Intelligence Peter Mika Yahoo! Research

Web 3 Peter Mika

Embed Size (px)

Citation preview

Page 1: Web 3 Peter Mika

Web 3.0 Combining Domain

Intelligence with Social Intelligence

Peter Mika

Yahoo! Research

Page 2: Web 3 Peter Mika

- 2 -

Two different aspects of intelligence

Intelligence

Domain Intelligence

Social Intelligence

Page 3: Web 3 Peter Mika

- 3 -

Domain Intelligence

• Goal is to teach machines how the world is put together

– Represent objects and relationships within a part of the world in a way that machines can act on this information

– Acting ~ reasoning. (Data integration is a simple form of reasoning. So is search.)

• Semantic Technologies

– RDF, OWL, SPARQL, HTTP etc.

• The Semantic Web is an application of semtech

– Publish data (Linked Data) and annotate textual resources (RDFa, microformats) using existing (shared) ontologies

– Semantic Search, Data Integration on a web scale

Page 4: Web 3 Peter Mika

- 4 -

Social Intelligence

• Goal is to teach machines how to manage and use human relationships, in particular apply a social filter to information and communication

– Turns out to be hard!

• Web 2.0 is a human-powered solution

– Very little automation: mostly helping humans to manage their networks a little more efficiently (invite 100 friends at the same time)

– Machine understanding is limited to “counting friends”, completing triangles

– In particular, little understanding of sharing through relationships

• In anything, information overload only got worse

Page 5: Web 3 Peter Mika

- 5 -

Web 3.0: combine domain and social intelligence

• Attempts so far at infusing semtech into Web 2.0 focused on solving the network portability issue (e.g. FOAF, dataportability.org, portable contacts etc.)

– Broke down on non-technical challenges, became less relevant

• Idea: better focus on content and relationships

– Opportunity: both media consumption and social interactions are moving online

– Models of how sharing, influence work could actually help address the overload of user-generated content

Page 6: Web 3 Peter Mika

- 6 -

Example: semantic bookmarking (prototype)

Page 7: Web 3 Peter Mika

- 7 -

Social Bookmarking

• Bookmark objects, instead of documents

• Connect users based on shared interests (aka object-centered sociability)

• See also Glue from AdaptiveBlue

• Why do it?

– Increase personal relevance

– Build rich “semantic” user profiles for content recommendation, semantic advertizing

Page 8: Web 3 Peter Mika

- 8 -

Example: Semantic Publishing (Zemanta Balloons)

Page 9: Web 3 Peter Mika

- 9 -

Semantic publishing

• Annotate content semantic tags (see commontag.org)

• Manual vs. automated solutions

• Why do it?

– Richer interlinking of content

– Direct interaction with the text, e.g. read more of this aspect of the story

– Richer user profiles

– Only solution for aggregating short forms of media

• Hashtags in Twitter

Page 10: Web 3 Peter Mika

- 10 -

Web 3.0

• To me, Web 3.0 has to be the semantic web, because it possibly holds the solution to at least some of the problems brought by Web 2.0

• Three steps

1. Infuse semantics into content

2. Observe users’ interactions with each other and the content

3. Build valuable insights on both individual media consumption and how human relationships function