Exploratory Search upon Semantically Described Web Data Sources: Service registration and methodology. At vldb2012

Exploratory Search upon Semantically Described

Web Data Sources

SSW workshop @ VLDB 2012, Istanbul, Turkey

Marco BrambillaPolitecnico di Milano

[email protected]

Outline

• Context

• Search Services Specification• Description• Semantic Annotation

• Exploratory Search

• Design patterns

• Demo

• Outlook

Context

Web is a huge, heterogeneous data source:

Structured, unstructured and semi-structured data Known problems of trust, reputation, consistency

User needs to solve real-life problems, not to find a web site

Context

Google? Well, yes… an “interesting” system

Context

User needs to solve real-life problems, not to find a web site

Web queries get increasingly complex and specialized Exploratory search From document search to object search

Search as a service

Viability of systems based upon search service orchestration

What are search services?

• APIs over Web data sources • Structured data• Domain-specific

• Wrapping of information utility sites

How can we use them?

“… search for upcoming concerts close to an attractive location (like a beach, lake, mountain, natural park, and so on), considering also availability of good, close-by hotels …”

• Applying complex queries (also with “joins”)

Background: semantic multi-domain search

“… expand the search to get information about available restaurants near the candidate concert locations, news associated to the event and possible options to combine further events …”

Liquid Query: Query Submission

Concert query conditions

Hotelsquery conditions

Example Scenario 1: Trip planner for events

Liquid Query: Query Execution

Liquid Query: alternative visualizations and domain-independent platform

Example Scenario 2: Scientific Publication search

Problem 1: Service specification

• No service description per se

• Focused on search• Ranking aware

• Description• Bottom-up• Based on the service interface

• Annotation• Relying on an external reference knowledge base

SDF and SAF

Service Description (SDF) vs. Service Annotation (SAF)

Example of SDF instances

The registration of services

Bottom-up approach from the service signatures

Registration process fully specified and implemented

See demo video at: http://search-computing.it/registration_demo

• Starts from SI details (name, type of service, etc.) and SI field details, i.e. name, data type and I/O directionality

• the name and I/O fields of the SI are scanned with NLP and Semantic techniques in order to identify the most suitable Domain Diagram items to represent them

• The expert user's intervention is required to provide a feedback concerning system-hypothesized mappings

• When all mappings have been validated, a newly created Access Pattern and its corresponding Service Mart are committed

Example of resulting service mart

• A set of predefined combinations of services, to be reused for specific cases

Problem 2: Reduce flexibility

Maximum flexibility over huge amounts of search services is not always the best solution

People want straightforward paths and want to be quick Commercial implementations are likely to be on fixed sets of domains and

fixed exploration directions

Design Patterns

• A set of blueprint combinations of services, to be reused for different cases

• Very much like UML design patterns or datamart patterns

Design Patterns – some examples

• Sequence

Design Patterns – some examples

• Star

Design Patterns – integrated examples

• Join

Design Patterns – integrated examples

• Join

Exploration implementation

• Not just a matter of data sources

• Also: data visualization, user interface specification, usability, ..

See demo videos at:http://demo.search-computing.net/night_planner_demo/seco/seco.htmlhttp://demo.search-computing.net/new_job_demo/seco/seco.html

Problem 3 - Outlook

When dealing with real-life problems, people do not trust the web completely

Want to go back to discussion with people Expect insights, opinions, reassurance

Exploratory search must be blended with social-network based recommendations and inputs

Social Search: increasing quality in search

• From exploratory search to friends and experts feedback

Exploratory Search System

Human Search System

Initial query

Exploration step

Exploration step

System API Social API

Database / IR index

Crowd / Community

Example: Find your job (social invitation)

Selected data items can be transferred to the crowd question

Find your job (response submission)

Conclusions and future work

Well, I’ve shown everything..

See our papers at WWW 2010 (Liquid Query) and WWW 2012 (CrowdSearcher)

Future work

• More experiments (e.g., vs. sociality of users, vs. crowds, …)

• Not only search: active integration of web structured data and social sensors

Some ads

• Search Computing book series (Springer LNCS)

• Workshop Very Large Data Search at VLDB

• VLDB Journal special issue (deadline Sept 2012)

Third book

coming

next Sept.

Marco Brambilla

Thanks!

Questions?

[email protected]