These are the slides from the tutorial on Semantic Digital Libraries we gave at International Conference for Digital Libraries and the Semantic Web (iCSD'09)
We need to introduce that Semantic Web is „a hype” – everyone is talking about – and say what it is
Then we have to show that actually the Web was meant to be that way from the begining
Finally tell what the „Semantic” means in all of that
OK – everyone says that he/she runs semantic web applications – but is that true?
Where are we *really* in the the SW layer cake?
Still far away – isn’t it ?
The semantic web will not emerge on its own
We need to help it
Explain the car example in more details
1)OK – first step to deliver semantic web – is to provide semantic descriptions
2)In order to do that – we can use widely addopted RDF
3)What is RDF? What are the properties?
4)RDF is not XML – as RDF is based on graph model
5)How machines can process it ?
Now, RDF is a really nice – but what more can we do with it?
First – obvious: objects/resources and relationsships
Than – easy extend the model with new relantionships, types, etc
Finally – we can easily integrate stuff
But be aware – RDF support many points of view – if you want to be sure – you need to exend the knowledge with more sentences to know which „conflicting” statement is right for you
RDF alone is not enough – it just delivers a graph – but we need to provide some constraints, knowledge, etc
Ontologies are the key concept to do so
They are based on the social agreement
There are several ways to express ontologies
The simplest ontology we can define is with RDF Schema
RDFS defines a small vocabulary ...
It can be easily used to define other vocabularies
Like in our example ...
Another example is OWL (motice that the name does not match the abbreviation – but we know why :D)
It is based on RDF/XML vocabulary
It allows us to be more explicite in what we define – see example ...
.. And DLs
OWL as W3C recommendation
We cannot see SW unless we will look into applications
There are two kinds:
enabling technologies like e.g.
End-user applications
Today we will present the new emerging ones – Semantic Digital Libraries
What are SDL?
They aim to integrate information coming from different sources
SDLs provide interoperability mechnisms – and can act as an integration champions in heterogeneous networks of libraries
But the **ultimate goal** is to provide better service for the end user – that is search and browsing features
One of the requirements was to support different metadata
Although they are so different – they can be represented in RDF
One of the requirements was to support different metadata
Although they are so different – they can be represented in RDF
One of the requirements was to support different metadata
Although they are so different – they can be represented in RDF
One of the requirements was to support different metadata
Although they are so different – they can be represented in RDF
The goal of this ontology is to provide means to mediate between heterogeneous database
schemas and metadata structures. The central concept of CIDOC is that knowledge is attained
by investigating relations between the facts. Therefore, the ontology concentrates on the
definition of relationships between items rather then the terminology of a particular domain.
The core ontology defines a set of very general classes (e.g. actor, event, period) and supplies a
variety of relationship descriptions that can adjust the concepts understanding for a particular
environment. Apart of the content preservation and summery CIDOC provides the concept
of events, such as creation, publication, etc. Therefore, with regard to digital libraries, it
is possible to utilize CIDOC to talk in detail about items content and their bibliographic
description.
1) Explain what do we mean by “simple solution” and “complex solution”
3 major problems o libraries:
They were build for librarians
They were delivering information – not sharing knowledge
DLs lost the human-part
Solutions:
Make users involved in the process
Allow users to share knowledge
Provide better communication means witin and outside of communities
Ok, we know what we are missing – what is the solution? Web 2.0 – a community-aware web
What is the focus of Web 2.0
What are in **general** tools technologies that we can call Web 2.0
Roadmap to SSIS
Roadmap to SSIS
Roadmap to SSIS
Roadmap to SSIS
Roadmap to SSIS
Roadmap to SSIS
Comparing different technologies
... And there are social semantic digital libraries finally :D
1)What is simile
2)Its relation to DSpace
3)Its goal
Main features of SIMILE:
Enhanced end-user services
Web-based architecture for digital assets dissimination
Now – lets see what SIMILE is
Two categories of components – that we will describe in details later
Now – lets see what SIMILE is
Two categories of components – that we will describe in details later
RDFizers -> in MarcOnt we would call them „adapters”
Solvent lets you easily integrate different sites into RDF -> e.g. For piggy bank
Now – lets see what SIMILE is
Two categories of components – that we will describe in details later
Themes - use cases; JHU collaboration for data curation
Encyclopedia of Chicago: integration of newspapers, maps, GIS, images; UVa Collectus tool
Max Planck - e-research; medical imaging
Max Planck has extended ontology; and PLoS
UVa - silos for arch drawings and letters - integrated
Integration of data sets with articles - data curation scenario
Are using triplestore to record assertions of annotations at fine-grained level. Use separate external triplestore. Exploding key metadata as RDF. Also assertions about agents - moving toward access control model. Mulgara - Topaz needs complex queries
Different triplestore than PLoS
How do we actually use it - how do we build on all these capabilities?
EV has two interrelated goals: provide a community discussion/presentation forum; create context for science resources in the library
Here’s a current example: adding polar bears to the endangered species list due to climate change. Note the NSDL tagged resource.
Search shows both the post and referenced resources now in the NSDL.
Walk through an example using MediaWiki. Create a new article.
And identify resources to include in the wiki article
Referenced resources display with NSDL icon
Contribute the wiki article to the library, adding metadata about the new resource.
And potentially add any referenced resources not in the library or more information about those that do exist.
- present overview and motivation behind JeromeDL
- describe ontologies and system architecture
- sebastian will take over
- after lunch you will have chance to get your hands on the newest version- let’s get started
- we’ve got phenomenal participation from users form all over the world and fantastic feedback
- users create their own customizations and at the same time influence the the main line of development
- here are the most important features of the system- to start with
- from the beginning our goal was create a semantic digital library with the user in the center
- we defined two questions in our research
- thus we tried to find the best way to integrate ...- interconnect
- we defined 3 different groups of users with 3 sets of requirements
- we can observe that those requirements are strongly based on the the legacy libraries
- in the quest of finding the answers to our question we knew that we need
- motivations based on the use case scenarios
- this is a three-layer architecture of metadata management on top of a digital library system- each layer enriches basic information gathered in a library with semantic annotations, thus providing additional capabilities to searching and browsing- the bottom layer handles tasks required from a digital object repository, that is keeps track of pysical representation of resources their structure and provenance- the middle layer lifts up egacy bibliographic descriptions to a semantic level- the top layer utilizes benefits from engging community of users into annotatiog and filtering resources
From alexandria DL -> to 3 layers -> to 3 metadata layers
- dynamic collections
MarcOnt Initative aims to utilize existing metadata in MarcOnt ontology through MMS
MarcOnt Initative aims to utilize existing metadata in MarcOnt ontology through MMS
From alexandria DL -> to 3 layers -> to 3 metadata layers
The underlying concepts of social networks can be also represented by RDF with FOAF metadata
We use FOAFRealm system to manage securely FOAF information, deliver authentification module and ... SSCF module
Now, legacy metadata is not enough – since we want average people to use our library
That is where SSCF comes into place
... And this is how it looks like
- let’s not take a look at ontologies used in JeromeDL
- structure ontology is used to describe an electronic representation of the content itself- information object - abstract library resource -, e.g. book, chapter, page- hasPart, hasParts- Resource Aggregation Service as a way for dynamic content- the upcoming verion of JeromeDL will combine RAS with OAI-ORI datamodel
- simplified view of the bibliographic ontology- keywords:
- domains:
- FOAFREalm ontoloy allows to capture user contribution to the resource descriptions
- FOAFRealm extends FOAF vocabulary with the notion of friendship level properties
- the community-aware ontology enables JeromeDL to describe social’s network information
- library resource as a bookmark (SSCF) - common practice of bookmarking,
- SSCF enables users to share their bookmarks with others, as well as annotate directories with keywords and domains- library resource as a Blog entry- users can comment a resource, thus providing new knowledge to the library
based on the tag ontology by Tom Gruber
tagging connects taggera, with document, with termg
community annotations for multimedia (currently in alpha stage)
the goal is to allow tagging in any type of documents
region of interest (ROI) tagging in photos
time-tagging of video streams
tags can contain descriptions, keywords, links and cross references
- when we will put all those ontologies together – this is what we got
What has Orang-utan from the Unseen University in Ankh-Morpork to do with semantic digital libraries?
Well – he was the first librarian to develop a very complicated controlled vocabullary for describing resources.
It consisted out of **one** word Ook – with various modifications like „Oook”, „gook”, „eek”, „eeek”.