Social Semantic Desktop Reference Architecture Evaluation

Post on 15-Jan-2016

28 views 0 download

description

Date. Social Semantic Desktop Reference Architecture Evaluation. People. Evaluation purpose (generic). To involve various stakeholders of the Social Semantic Desktop community in the critical assessment of the Social Semantic Desktop Blueprint - PowerPoint PPT Presentation

Transcript of Social Semantic Desktop Reference Architecture Evaluation

© NEPOMUK Consortium 2006 –

2008

Integrated Project n°27705 – Priority 2.4.7 – Semantic knowledge based

systems

Social Semantic DesktopReference Architecture

EvaluationDate

People

Reference Architecture Evaluation

2Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• To involve various stakeholders of the Social Semantic Desktop community in the critical assessment of the Social Semantic Desktop Blueprint

• Refine the scope functional boundaries of the Social Semantic Desktop Blueprint

• To involve the stakeholders in the architectural design by bringing their experience and expertise

Evaluation purpose (generic)

Reference Architecture Evaluation

3Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Evaluation process

• Preparation• Stakeholders profile – Questionnaire A (10 min.)• Introduction (5 min.)• Motivation and Overview (2 min.)• The Social Semantic Desktop Blueprint (40 min.)

• Analysis• Scenarios for the Social Semantic Desktop Blueprint (5

min.)• Interactive scenarios revision (20 min.)

• Synthesis• Assessment – Questionnaire B (10 min.)• Final assessment – processing the evaluation data to

formulate the evaluation results

Reference Architecture Evaluation

4Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Please fill-in Questionnaire A …

http://www.surveymonkey.com/s.aspx?sm=Y7dGTHTbWa_2biudS_2fM8924g_3d_3d

Reference Architecture Evaluation

5Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

The Social Semantic Desktop

• Extension of the personal desktop …

• … into a collaboration environment

• Goals:• Improve personal information management

• Improve cross-media and cross-application linking

• Improve sharing and exchange across social and organizational relations.

Reference Architecture Evaluation

6Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Personal Information Management

•Distributed Information Management

•Social Networks and Community Services.

Social Semantic Desktop Layers

Reference Architecture Evaluation

7Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

SSD evolution

Reference Architecture Evaluation

8Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

SSD evolution (cont.)

Reference Architecture Evaluation

9Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

SSD evolution (cont.)

Reference Architecture Evaluation

10Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

SSD evolution (cont.)

Reference Architecture Evaluation

11Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

SSD evolution (cont.)

Reference Architecture Evaluation

12Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

SSD evolution – current status

Reference Architecture Evaluation

13Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

The NEPOMUK IP Project

•FP6 Project IST

•17 Partners:• 8 research centres• 4 big industry

players• 3 SMEs

•Open source

Reference Architecture Evaluation

14Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Goals

• Definition of the Social Semantic Desktop Blueprint

• Standardization of ontologies and APIs

• Development• Prototypes• The reference Social Semantic Desktop

implementation

The NEPOMUK IP Project (cont.)

Reference Architecture Evaluation

15Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Evaluation purpose (technical aspect)

Reference Architecture Evaluation

16Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Social Semantic Desktop Blueprint• Personal information management -> Personal

knowledge creation and organization

• Data interoperability -> Cross-media and cross-application linking

• (Distributed) Social collaboration -> Sharing, exchange and alignment of the personal knowledge in a distributed manner

• Evaluation = Proof that the SSD Blueprint is able to handle all the possible scenarios arising in the space of these three dimensions

Evaluation purpose (technical aspect)

© NEPOMUK Consortium 2006 –

2008

Integrated Project n°27705 – Priority 2.4.7 – Semantic knowledge based

systems

The Social Semantic Desktop Blueprint

Reference Architecture Evaluation

18Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Overview

•Social Semantic Desktop – Scenario

•Social Semantic Desktop – Engineering cycle

•Social Semantic Desktop Models

•The Social Semantic Desktop Blueprint

Reference Architecture Evaluation

19Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Scenario

Reference Architecture Evaluation

20Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Engineering cycle

Reference Architecture Evaluation

21Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Engineering cycle (I)

Reference Architecture Evaluation

22Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Usage – Scenarios – Functionalities

Reference Architecture Evaluation

23Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Engineering cycle (II)

Reference Architecture Evaluation

24Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Knowledge Articulation and Visualization• Semantic data editing and presentation

• Standard Desktop Classification Structures• Standard set of vocabularies and ontologies (e.g.

calendar, task management)• Mapping and Aligning Algorithms

• Alignment of information from similar domains expressed with different schemas

• Wrapping of Legacy Information• Standardized semantic representation of structured

and unstructured data

Technical requirements (I)

Reference Architecture Evaluation

25Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Metadata Storage and Querying• Central place for storing and querying the information

and the associated metadata• Linking of Data Items and Relational

Metadata• Link of arbitrary information across different media

types, file formats and applications• Social Aspects

• Social relation building and knowledge sharing within social communities

• Open Architecture• Clearly defined and published interfaces• Open for integration with external adopters

Technical requirements (II)

Reference Architecture Evaluation

26Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Engineering cycle (III)

Reference Architecture Evaluation

27Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Models

Reference Architecture Evaluation

29Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Personal Information Model (PIM)• Vocabulary allowing individual persons to express

their own mental models in a structured way• Different mental models can be integrated based on

matching algorithms or on domain ontologies

• Information Element Model (IEM)• Vocabulary for describing information elements which

are commonly present on the semantic desktop 

• Annotation Model (AM)• Vocabulary, commonly required to annotate resources

on the semantic desktop

Models

Reference Architecture Evaluation

30Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – PIMO

Reference Architecture Evaluation

31Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – PIMO

Reference Architecture Evaluation

32Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Requirements• a representation of abstract concepts: Love, Rome, Acme Inc. • a representation of concrete, addressable resources: "w3c

homepage at www.w3.org" • a representation of documents: "the document at

http://www.w3.org/" • multiple names for a thing: "Love", "Liebe“; "W3", "WWW" • same name for two different things: "Apache - helicopter",

"Apache - software". • class-subclass relations: a subclass has all properties of the

superclass + its own • class-instance relations • part-of relations: the city of Rome is part of Italy • related information: Spaghetti is related to Italy • data properties to describe details: Rome has a population of

2.8 mio • document-has-topic: the document

"http://www.w3.org/2001/sw" is about the "Semantic Web" • a representation of time: the document was created in 2005.

The project started on 1.1.2006

NEPOMUK Models – PIMO

Reference Architecture Evaluation

34Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•The NEPOMUK Information Element (NIE)

• Set of ontologies

• Vocabulary for describing information elements commonly present on the semantic desktop

NEPOMUK Models

Reference Architecture Evaluation

35Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•NIE Core – NEPOMUK Information Element Core Ontology

•NFO – NEPOMUK File Ontology•NCO – NEPOMUK Contact Ontology•NMO – NEPOMUK Message Ontology•NCAL – NEPOMUK Calendar Ontology•NEXIF – NEPOMUK EXIF Ontology•NID3 – NEPOMUK ID3 Ontology

NEPOMUK Models – NIE

Reference Architecture Evaluation

36Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NIE

Reference Architecture Evaluation

37Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NIE Core

Reference Architecture Evaluation

38Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NFO

Reference Architecture Evaluation

39Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NMO

Reference Architecture Evaluation

40Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NCAL

Reference Architecture Evaluation

41Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NCO

Reference Architecture Evaluation

42Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• The NEPOMUK Annotation Ontology (NAO)• Vocabulary that enables users to attach custom

descriptions, identifiers, tags and ratings to resources on their desktop

• Via other properties, the user is also able to make generic relationships between related resources explicit.

• Relationships between resources that are too general to be included at the domain ontology level are also defined in the annotation ontology

• Given the high-level status of this ontology, these properties can be used to link any related resources on the user's desktop, as well as provide custom human-readable textual annotations

NEPOMUK Models

Reference Architecture Evaluation

43Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

NEPOMUK Models – NAO

Basic

Specific

Conventional tagging

Reference Architecture Evaluation

44Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint

Reference Architecture Evaluation

45Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint

Reference Architecture Evaluation

46Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint

Reference Architecture Evaluation

47Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint

Reference Architecture Evaluation

48Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint (design rationale)

Reference Architecture Evaluation

49Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint (design rationale)

Reference Architecture Evaluation

50Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint – Social Services

Reference Architecture Evaluation

51Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Foundational layer for achieving social collaboration

•Main functionality

• Distributed information management

• Communication

• Security

Social Services

Reference Architecture Evaluation

52Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Enables cross-desktop communication

•Handles both service-to-service and human-to-human communication

•Acts as message-carrier for the Notification service

Messaging service

Reference Architecture Evaluation

53Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Enables seamless management of groups of persons

•Should interact closely with the Messaging service and the Access Control Management

Group management

Reference Architecture Evaluation

54Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Defines and applies (fine-grained) access control on the resources present in the shared information space

•Should interact closely with the Group Management

Access control management

Reference Architecture Evaluation

55Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Represents the main access point to the social services infrastructure

•Should ensure a transparent single-sign-on process

Authentication

Reference Architecture Evaluation

56Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Represents the backbone of the Social Services layer

•Insures proper distributed storage and search of the resources present in the shared information space

•Should support in a transparent way both data and metadata

Distributed storage and search

Reference Architecture Evaluation

57Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Overall interaction recommendation

OROR

Reference Architecture Evaluation

58Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint – Semantic Desktop Services

Reference Architecture Evaluation

59Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Semantic Desktop middleware – foundational layer for achieving

• Personal Information Management• Data interoperability

• Structure• Core services – minimum set of services needed to have a

Semantic Desktop

• Data services – additional set of services needed to achieve data interoperability

• Other – extra set of services targeting different particular functionalities

• Service registry – main access point to the semantic desktop middleware

Semantic Desktop Services

Reference Architecture Evaluation

60Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Core services

Reference Architecture Evaluation

61Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Central metadata and structured data store

• RDF Storage and access to semantic resources and their descriptions

• Provides full text indexing and inference

• Support for multiple query languages

• Main hub for data integration• Lifted data from the Data Wrapper and other services• User’s Personal Information Model

Local storage

Reference Architecture Evaluation

62Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• High-level wrapper for the PIM concepts present in the Local Storage

• Implements commonly used methods to manipulate a PIM

• Fully dependent on the Local Storage solution

• Typical functionalities:• creating and deleting classes, properties, and

resources • adding and removing ontologies • various convenience methods • the URI identification of the user • the personal namespace of the user • the URI of the user's personal information model

PIM Service

Reference Architecture Evaluation

63Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Lifting the data present on the desktop

• Semantic wrapper for the data (RDF conversion)

• Uses plug-ins to interact with legacy systems – each plug-in is specialised on extracting the data from one application format (such as the addressbook) or a file-type (such as PDF, MS-Word)

• Stores the RDF into the Local Storage

Data wrapper

Reference Architecture Evaluation

64Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Providing an abstraction layer for accessing the Local Storage and possibly other service which may be involved in the search process

• Enables the community to build arbitrarily complex extensions to the query processing workflow, without changing the corresponding API.

• Support for arbitrary query languages (language can be specified as a parameter to the query)

• Support for full-text search and ranking

• Support for search shortcuts (similar to stored procedures) of more complex SPARQL queries (e.g., my recent documents)

Local search

Reference Architecture Evaluation

65Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Generic service for multiple notification mechanisms

• Represents a glue between the services that perform tasks based on the results of the behaviour of other services

• Some notifications can be directly translated into user feed-back

• Examples:• Black-board• Publish-subscribe

Notification

Reference Architecture Evaluation

66Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Data services

Reference Architecture Evaluation

67Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Provides the mechanisms for:

• Finding mappings between ontologies

• Learning or enhancing mappings

•High-level information integration

•Querying data sources based on the learned mappings

•Examples: mappings between 2 PIMs

Social data alignment

Reference Architecture Evaluation

68Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• Aligns data present in the Local Storage

• Suggests new links or annotations between information elements and resources

• Finds entities that are usable as entries in the user's PIM

• Unifies multiple representations of the same entity into one thing (such as two information elements representing the same person into one)

• Improves results based on learning from the user’s feed-back

Local data alignment

Reference Architecture Evaluation

69Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Translating an RDF graph from one ontology language into another

•Translation from the languages used for manipulating the data present in the Local Storage is other formats

• Example: RDF graphs from FOAF to vCard

Data translation

Reference Architecture Evaluation

70Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Recommends similar or related resources as well as additional resource descriptions to a particular resource

•Based on the metadata describing a resource other resources and/or additional metadata descriptions are recommended

•The recommendations are can be performed not only from the Local Storage, but also from the Shared Information Space

Resource and metadata recommender

Reference Architecture Evaluation

71Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Other

Reference Architecture Evaluation

72Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Generic service for performing specific reasoning tasks

•Can be used for example by the recommendation services, or other services to reason on the data present in the Local Storage

Reasoning service

Reference Architecture Evaluation

73Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Creates abstract contexts and detects the current context by observing the user activity

•Should contain a user observation (logging) hub

•Generates context models to be used by the other services for learning or profiling

Context elicitation

Reference Architecture Evaluation

74Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Umbrella of services performing low-level natural language processing

•The results can be used for enriching the models of the information elements (lifting data)

•Examples• Keyword extraction• Speech acts detection• Rhetorical elements and relations extraction

Text analytics services group

Reference Architecture Evaluation

75Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Umbrella of specific services that compute recommendations for information elements, based on different criteria

•Examples:• Tag recommender• Community (group) recommender• Person recommender

Recommendation services group

Reference Architecture Evaluation

76Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Service registry

Reference Architecture Evaluation

77Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Central access point to all the services present in the platform

•Allows publishing, removing and discovery of services

•Represents a wrapper of the Semantic Desktop middleware

Service registry

Reference Architecture Evaluation

78Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Blueprint – Extensions and Presentation Layer

Reference Architecture Evaluation

79Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Services developed for specific needs

•Combine directly the functionality provided by the Semantic Desktop Middleware and the Social Services

•Can provide complex additional functionalities for the presentation layer

Semantic Desktop extensions

Reference Architecture Evaluation

80Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

• High level personal information management extension with focus on tasks

• Ad-hoc task planning and flexible changes

• Collaborative work on tasks

• Connecting tasks with personal models as well as group information objects, and integration into organizational processes

• Task patterns recognition and learning

Task Management extension

Reference Architecture Evaluation

81Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

•Knowledge workbench ~ ordinary day-to-day applications

•Integrate functionality provided by the layers below via plug-ins or add-ons

•Seamless integration of semantic features into the knowledge workbench

Presentation layer

Reference Architecture Evaluation

82Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Scenario - revisited

Reference Architecture Evaluation

83Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Scenario revisited – implementation

© NEPOMUK Consortium 2006 –

2008

Integrated Project n°27705 – Priority 2.4.7 – Semantic knowledge based

systems

Analysis

Reference Architecture Evaluation

85Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Objectives

•Is the current list of scenarios complete?

• Are there scenarios missing?• If yes, is the blueprint able to handle them?

•Does the blueprint support the revised list of scenarios?

Reference Architecture Evaluation

86Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

1. There should be a common framework for the customization and personalization of the desktop

2. There must be support of a single authentication step as a prerequisite for collaboration and access to the share information space

3. Access to the shared information space should be in a transparent manner

4. Collaboration should be transparently supported by context awareness

5. Document management and sharing should be in a transparent manner

6. Creation and management of collaborative groups should be transparent, supported by a simple management of people competence profile

7. Communication within a team should be automated supported by presence and context awareness mechanisms

8. There should be persistency of personal work spheres such that there is continuity between different working sessions

9. The document management of the work space should be semantically enriched, thus enabling sophisticated searches and filtering

Scenarios

Reference Architecture Evaluation

87Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Desktop customization and personalization (1)

• (US) When Claudia starts her SSD she wants to see her desktop customized/organized according to her user profile, context, user preferences, interactions, collaborations and tasks she is involved in.

• (SS) System should be able to authenticate user based on his credentials. Once authentication is done, system should be able to process user context to determine the activity space user wants to login to. If there are specific user preferences to personalize the user desktop then they should be processed to customize the user desktop. The system should also be able to process the user context to retrieve all the files/folders owning to the user context.

Reference Architecture Evaluation

88Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Single authentication (2)

• (US) Claudia is working in two different projects, with two slightly overlapping teams. Claudia wants to get simple access to sensitive data subject to strict security policy enforced in the two projects. She wants to be able to do this using a single login for all her profiles, involved in the different projects.

• (SS) The system will need to provide identity management for different profiles of the same user. The system will also need to provide an authentication mechanism to authenticate the profiles in the different shared information spaces. The shared information space system checks the user identity and in case of success, allows access to private data.

• Note: the same requirements are also relevant for the transparent access to the shared information space scenario

Reference Architecture Evaluation

89Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Transparent collaboration based on context awareness (4)

• (US) Claudia wants to search and share specific project or group resources (based on the working context and document relationships) from different projects into a particular share information space (created for example for the company’s audit). Claudia also wants to build an ad-hoc team based on the working context (with the other project managers).

• (SS) The system should allow to manage (i.e. create or update) the working context for the resource (e.g. documents or folder). The system should be able to process the current working context to search the resources related to the particular working context. The system should allow uploading and sharing of documents. Sharing involves also sending notifications to relevant users.

• System should provide a mechanism to handle data interoperability issues while sharing different data objects.

Reference Architecture Evaluation

90Social Semantic Desktop© NEPOMUK Consortium 2006 –

2008

Please fill-in Questionnaire B …

http://www.surveymonkey.com/s.aspx?sm=61wctkDVyCsCDVx3xaCzEw_3d_3d

© NEPOMUK Consortium 2006 –

2008

Integrated Project n°27705 – Priority 2.4.7 – Semantic knowledge based

systems

Thank you!