Тексты на английском
<<  The orange revolution and democratic change in Ucraine Ханс кристиан андерсен проект для  >>
Visualization, Triples, Serialization
Visualization, Triples, Serialization
What is Change
What is Change
What is Change
What is Change
What is Change
What is Change
What is Change
What is Change
Low-Level Delta
Low-Level Delta
Detection Algorithm for L (1/2)
Detection Algorithm for L (1/2)
Detection Algorithm for L (1/2)
Detection Algorithm for L (1/2)
Performance
Performance
Evaluation: Usefulness and Intuitiveness
Evaluation: Usefulness and Intuitiveness
Evaluation: Conciseness
Evaluation: Conciseness
Картинки из презентации «High-Level Change Detection in the Semantic Web» к уроку английского языка на тему «Тексты на английском»

Автор: Giorgos Flouris. Чтобы познакомиться с картинкой полного размера, нажмите на её эскиз. Чтобы можно было использовать все картинки для урока английского языка, скачайте бесплатно презентацию «High-Level Change Detection in the Semantic Web.ppt» со всеми картинками в zip-архиве размером 414 КБ.

High-Level Change Detection in the Semantic Web

содержание презентации «High-Level Change Detection in the Semantic Web.ppt»
Сл Текст Сл Текст
1High-Level Change Detection in the 19versions in a concise, intuitive and
Semantic Web. Giorgos Flouris correct way Low-level deltas Easier to get
fgeo@ics.forth.gr. Institute of Computer High-level deltas More concise (e.g.,
Science Foundation for Research and Rename_Class) More intuitive (e.g.,
Technology – Hellas Heraklion, Greece. Pull_Up_Class) Carry additional
Joint work with: Vicky Papavassiliou, information (e.g., Generalize_Domain)
Irini Fundulaki, Dimitris Kotzinos, Objective: detection of high-level deltas.
Vassilis Christophides. 16/09/2009. 16/09/2009. Giorgos Flouris. 19.
Giorgos Flouris. 1. 20Language of Changes and Algorithm.
2World Wide Web. WWW (and HTML) focus Deltas based on some language of changes A
on human readability Page presentation set of formal definitions that describe
(fonts, colors, images, …) Human the changes that can be understood and
understanding Presentation ? Semantical detected Can be high-level or low-level
content Content is not formally described Must be coupled with a corresponding
(for a machine to understand) WWW contains detection algorithm Low-level languages
documents, not data. 16/09/2009. Giorgos easy to define (Add(t), Del(t)) High-level
Flouris. 2. languages more complicated Several
3Problems with Current Web. Search and proposals; no standard Challenges for
access becomes difficult Software ignorant high-level languages Must be deterministic
of the semantical content of a web page (exactly one high-level delta) Must be
Keyword search High recall, low precision fine-grained enough to capture subtle
Terminological issues Synonyms (heart changes Must be coarse-grained enough to
disease = cardiac disease) be concise. 16/09/2009. Giorgos Flouris.
Hyponyms/hypernyms (parliament members are 20.
politicians) Queries on the semantical 21Proposed Language L. The formal
content cannot be made Fetch articles that definition of a change consists of:
support B. Obama’s foreign policy Fetch Changes required in the low-level delta
the home pages of all members of the Greek (added/deleted triples) Conditions that
Parliament. 16/09/2009. Giorgos Flouris. should hold in V1 and/or V2
3. Generalize_Domain(P, X, Y) Del([P domain
4Semantic Web. The Semantic Web is an X]) Add([P domain Y]) P existing property
extension of the current web in which in both V1, V2 X, Y existing classes in
information is given well-defined meaning, both V1, V2 X subclass of Y in both V1, V2
better enabling computers and people to Generalize_Domain(participants, Onset,
work in cooperation (Berners-Lee et al., Event): detectable Similarly for the other
2001) The Semantic Web provides a common changes in L (about 120 in total).
framework that allows data to be shared 16/09/2009. Giorgos Flouris. 21.
and reused across application, enterprise, 22Results on L: Granularity. Granularity
and community boundaries problem: solved by defining levels of
http://www.w3.org/2001/sw/ [Semantic Web] changes Basic Changes: fine-grained,
is a collaborative effort led by W3C with roughly correspond to low-level Composite
participation from a large number of Changes: coarse-grained, group several
researchers and industrial partners basic changes together Heuristic Changes:
http://www.w3.org/2001/sw/. 16/09/2009. based on heuristics, necessary for Rename,
Giorgos Flouris. 4. Merge, Split etc Problems with determinism
5Semantic Web in Practice. Web of data, One evolution could correspond to
rather than documents HTML for different sets of basic/composite changes
presentation Semantical languages for Priorities in detection Heuristic ?
semantical content Readable and Composite ? Basic. 16/09/2009. Giorgos
understandable by humans and machines Flouris. 22.
Semantic Web languages, protocols, etc Web 23Results on L: Types of Changes.
page annotation (metadata descriptions Changes. Low-Level. High-Level. Add Del.
etc) Publication of data on the Internet Basic. Composite. Heuristic.
Efficient communication and manipulation Delete_Subclass Delete_Domain.
of data over the Internet Different Pull_Up_Class Change_Domain. Rename_Class
applications Efficient searching Sharing Split_Class. 16/09/2009. Giorgos Flouris.
of data (e-science, e-government, remote 23.
learning, …). 16/09/2009. Giorgos Flouris. 24Results on L: Determinism. Each
5. low-level change is associated with
6Ontologies. Backbone of the Semantic exactly one detectable high-level change
Web Ontologies allow the description of Full partitioning of low-level changes
data Annotation and metadata regarding web into high-level ones Each pair of versions
pages Terminological relations (synonyms, (V1, V2) is associated with: Exactly one
hyponyms, …) Communication and description low-level delta Exactly one high-level
of data, ideas, beliefs An ontology is an delta Determinism is necessary More than
explicit specification of a shared one would lead to ambiguities Less than
conceptualization of a domain (Gruber, one would make some inputs (V1, V2)
1993) Precise, logical account of the irresolvable. 16/09/2009. Giorgos Flouris.
intended meaning of terms, data structures 24.
etc Common (shared) interpretation of 25Results on L: Application. Version 1
terms Formal vocabulary for information (V1). Version 2 (V2). Detect C. Apply C.
exchange (for humans and machines). Apply C-1. Period. participants. Actor.
16/09/2009. Giorgos Flouris. 6. Event. Actor. Event. started_on. Birth.
7Ontologies in Practice. Basic Persistent. Onset. participants.
structures: Classes (or concepts): started_on. Onset. Existing. Stuff. Stuff.
collections of objects (e.g., Actor, Birth. participants. G_Birth. Giorgos.
Politician) Properties (or roles): binary participants. Giorgos. G_Birth.
relationships between objects (e.g., 16/09/2009. Giorgos Flouris. 25.
started_on, member_of) Instances (or 26Results on L: Deltas Keep Version
individuals): objects (e.g., Giorgos, B. History. Can reproduce all versions as
Obama) Relations between them Subsumption long as you keep (any) one version and the
(Parliament_Member subclass of deltas Deltas are more concise than the
Politician), instantiation (B. Obama versions themselves Storage and
instance of Politician), … The allowed communication efficiency. 16/09/2009.
relations and their semantics depend on Giorgos Flouris. 26.
the language Different representation 27Detection Algorithm for L (1/2). Run
languages for ontologies RDF, RDFS, Matcher (External). Compute Heuristic
DAML+OiL, OWL, OWL-DL, OWL-Lite, OWL2, Changes. Calculate Low-Level Delta. List
DLs, … Usually triple-based. 16/09/2009. of Mappings <V1:Existing> is matched
Giorgos Flouris. 7. with <V2:Persistent> Heuristic
8Visualization, Triples, Serialization. Changes Rename_Class(Existing,
Visualization. Triple Representation. Persistent). Triples in Delta (step 1:
Serialization (RDF/XML). Define classes low-level) Del([participants domain
[Period type Class] Define properties Onset]) Del([Birth subclass Onset])
[participants type Property] [participants Del([Event subclass Period]) Del([Existing
domain Onset] [participants range Actor] type Class]) Del([Stuff subclass
Instantiate/define individuals [G_Birth Existing]) Del([started_on domain
type Birth] [Giorgos type Actor] [G_Birth Existing]) Del([Period type Class])
participants Giorgos] Define hierarchies Add([Birth subclass Event])
[Event subClass Period]. Period. Actor. Add([participants domain Event])
Event. participants. started_on. Onset. Add([Persistent type Class]) Add([Stuff
Existing. Stuff. Birth. participants. subclass Persistent]) Add([started_on
Giorgos. G_Birth. <rdfs:Class domain Persistent]). Triples in V2
rdf:ID=“Period”> </rdfs:Class> (Partial List) [Event type Class]
<rdf:Property rdf:ID=“participants”> [participants type Property] [Event domain
<rdfs:domain rdf:resource=“Onset”/> participants] [participants range Actor]
<rdfs:range rdf:resource=“Actor”/> [Giorgos type Actor] [Persistent type
</rdf:Property> <G_Birth Class] [Stuff subclass Persistent]
rdf:about Birth> <participants> [started_on domain Persistent] [Onset
<Giorgos rdf:about Actor/> subclass Event] [Birth subclass Event] …
</participants> </G_Birth> Triples in V1 (Partial List) [Period type
<rdfs:Class rdf:ID=“Event”> Class] [Event subclass Period]
<rdfs:subClassOf [participants type Property] [participants
rdf:resource=“Period”/> domain Onset] [participants range Actor]
</rdfs:Class> 16/09/2009. Giorgos [Existing type Class] [Stuff subclass
Flouris. 8. Existing] [started_on domain Existing]
9Ontology Dynamics. Ontologies change [Onset subclass Event] … 16/09/2009.
constantly World changes (dynamic models) Giorgos Flouris. 27.
View on the world changes (new knowledge, 28Detection Algorithm for L (2/2). ? ? ?
measurements, etc) Perspective and usage Find Associated Change. Del([participants
changes Example: GO ontology changes daily domain Onset]).
Gene Ontology: information about gene Generalize_Domain(participants, Onset,
products (biology) Must find a way to cope Event) DETECTABLE. Triples in V2 (Partial
with changes Ontology evolution (modify an List) [Event type Class] [participants
ontology in response to a change) Ontology type Property] [Event domain participants]
versioning (keep track of versions and [participants range Actor] [Giorgos type
their relations) … We deal with a Actor] [Persistent type Class] [Stuff
peripheral problem (change detection). subclass Persistent] [started_on domain
16/09/2009. Giorgos Flouris. 9. Persistent] [Onset subclass Event] [Birth
10What is Change? Real World. Ontology subclass Event] … Triples in V1 (Partial
Evolution Algorithm. Delete_Class(…) List) [Period type Class] [Event subclass
Pull_Up_Class(…) Rename_Class(…) … Period] [participants type Property]
Ontology. 16/09/2009. Giorgos Flouris. 10. [participants domain Onset] [participants
11What is Change Detection? Real World. range Actor] [Existing type Class] [Stuff
Change Detection Algorithm. subclass Existing] [started_on domain
Delete_Class(…) Pull_Up_Class(…) Existing] [Onset subclass Event] … Triples
Rename_Class(…) … Ontology. 16/09/2009. in Delta (step 2: heuristic)
Giorgos Flouris. 11. Del([participants domain Onset])
12Keeping Track of Changes. Purpose of Del([Birth subclass Onset]) Del([Event
this work: change detection A posteriori subclass Period]) Del([Period type Class])
detect the differences (delta or diff) Add([Birth subclass Event])
between versions in a concise, intuitive Add([participants domain Event])
and correct way It is important to store Rename_Class(Existing, Persistent).
the changes between versions Visualization Triples in Delta (step 3: basic and
of differences Efficient storage and/or composite) Del([Birth subclass Onset])
communication Evolution history Record Del([Event subclass Period]) Del([Period
changes as they happen (manual or type Class]) Add([Birth subclass Event])
automatic) Error-prone, difficult (often Rename_Class(Existing, Persistent)
impossible). 16/09/2009. Giorgos Flouris. Generalize_Domain(participants, Onset,
12. Event). Triples in Delta (step 4: result)
13Sample Evolution. Version 1 (V1). Delete_Class(Period, ?, {Event}, ?, ?, ?,
Version 2 (V2). Evolution. Period. ?) Pull_Up_Class(Birth, Onset, Event)
participants. Actor. Event. Actor. Event. Rename_Class(Existing, Persistent)
started_on. Birth. Persistent. Onset. Generalize_Domain(participants, Onset,
participants. started_on. Onset. Existing. Event). 16/09/2009. Giorgos Flouris. 28.
Stuff. Stuff. Birth. participants. 29Find Associated Change.
G_Birth. Giorgos. participants. Giorgos. Del([participants domain Onset]).
G_Birth. 16/09/2009. Giorgos Flouris. 13. Del([participants domain Onset]). Required
14Analyzing the Evolution (Using in Low-Level Delta. Potentially Associated
Triples). Triples in V1 (partial list) High-Level Change. Add([participants
[Event type Class] [Period type Class] domain X]).
[Event subclass Period] [participants type Generalize_Domain(participants, Onset, X).
Property] [participants domain Onset] Add([participants domain X]).
[participants range Actor] [Giorgos type Specialize_Domain(participants, Onset, X).
Actor] [Existing type Class] [Stuff ---. Delete_Domain(participants, Onset).
subclass Existing] [started_on domain Del([participants type Property])
Existing] [Onset subclass Event] [Birth Del([participants range X]).
subclass Onset] … Triples in V2 (partial Delete_Property(participants, Onset, X). …
list) [Event type Class] [participants … Operations Pull_Up_Class(*,*,*) [not in
type Property] [Event domain participants] the table]
[participants range Actor] [Giorgos type Delete_Property(participants,*,*)
Actor] [Persistent type Class] [Stuff [necessary triples not found]
subclass Persistent] [started_on domain Specialize_Domain(participants, Onset,
Persistent] [Onset subclass Event] [Birth Event) [conditions not true]
subclass Event] … 16/09/2009. Giorgos Generalize_Domain(participants, Onset,
Flouris. 14. Birth) [wrong parameter (triples not
15Low-Level Delta. Triples in V2 but not found)] Generalize_Domain(participants,
in V1 (added triples) [Event domain Onset, Event) [DETECTABLE (ASSOCIATED)]
participants] [Persistent type Class] Delete_Domain(participants, Onset)
[Stuff subclass Persistent] [started_on [composite changes have priority].
domain Persistent] [Birth subclass Event]. 16/09/2009. Giorgos Flouris. 29.
Triples in V1 but not in V2 (deleted 30Implementation. Algorithm implemented
triples) [Period type Class] [Event for experiments and evaluation Uses the
subclass Period] [participants domain APIs of SWKM Platform for efficient and
Onset] [Existing type Class] [Stuff scalable management of dynamic RDF/S
subclass Existing] [started_on domain ontologies and data Query, update,
Existing] [Birth subclass Onset]. low-level delta, high-level delta,
Low-Level Delta Add([Event domain versioning, … 16/09/2009. Giorgos Flouris.
participants]) Add([Persistent type 30.
Class]) … Del([Period type Class]) … 31Performance. Complexity:
16/09/2009. Giorgos Flouris. 15. O(max{N1,N2,N2}) Linear average-case
16Analyzing the Evolution (Visually). Highly dependent on the detected changes
Version 1 (V1). Version 2 (V2). Evolution. (type, number). 16/09/2009. Giorgos
High-Level Delta Flouris. 31.
Generalize_Domain(participants, Onset, 32Evaluation: Usefulness and
Event) Pull_Up_Class(Birth, Onset, Event) Intuitiveness. L is well-defined (changes
Delete_Class(Period, ?, {Event}, ?, ?, ?, used in practice) GO: add/delete class,
?) Rename_Class(Existing, Persistent). comments changing CIDOC: add/delete/rename
Period. participants. Actor. Event. properties Results confirmed by
started_on. Actor. Event. Birth. literature/editor notes. 16/09/2009.
Persistent. Onset. participants. Giorgos Flouris. 32.
started_on. Onset. Existing. Stuff. 33Evaluation: Conciseness. Basic ?
participants. G_Birth. Giorgos. Stuff. Low-Level Basic+Composite+Heuristic
Birth. participants. Giorgos. G_Birth. << Low-Level. 16/09/2009. Giorgos
16/09/2009. Giorgos Flouris. 16. Flouris. 33.
17Comparing the Deltas. Version 1 (V1). 34Manual Change Recording (CIDOC).
Version 2 (V2). Evolution. Low-level Editor notes Delete class: 3 Add property:
delta. High-level delta. Period. 54 Delete property: 16 Rename property: 24
participants. Actor. Event. started_on. Redirect properties (domain): 14 Redirect
Actor. Event. Birth. Persistent. Onset. properties (range): 14. Detection result
participants. started_on. Onset. Existing. Delete class: 6 Add property: 58 Delete
Stuff. participants. G_Birth. Giorgos. property: 18 Rename property: 30
Stuff. Birth. participants. Giorgos. Generalize_Domain: 13 Specialize_Domain: 1
G_Birth. 16/09/2009. Giorgos Flouris. 17. Generalize_Range: 14 Specialize_Range: 1
18Associations (Partitioning). Low-Level Change_Range: 1. 16/09/2009. Giorgos
Changes. Associated High-Level Changes. Flouris. 34.
Del([participants domain Onset]). 35Conclusion. High-level change
Generalize_Domain (participants, Onset, detection A posteriori detection (input:
Event). Generalize_Domain (participants, V1, V2) No further information needed
Onset, Event). Add([participants domain (e.g., logs, change recording etc) Formal
Event]). Del([Birth subclass Onset]). semantics Formal results (reversibility,
Pull_Up_Class(Birth, Onset, Event). determinism, …) Non-heuristic based
Pull_Up_Class(Birth, Onset, Event). (except for heuristic changes) No need for
Add([Birth subclass Event]). Del([Period precision and recall evaluation Efficient,
type Class]). Delete_Class (Period, ?, sound and complete detection algorithm
{Event}, ?, ?, ?, ?). Delete_Class Nice informal properties Conciseness,
(Period, ?, {Event}, ?, ?, ?, ?). intuitiveness Future work: more
Del([Event subclass Period]). operations, evaluation on other datasets,
Del([Existing type Class]). evaluation with real users. 16/09/2009.
Rename_Class(Existing, Persistent). Giorgos Flouris. 35.
Rename_Class(Existing, Persistent). 36References. Thank You. Vicky
Rename_Class(Existing, Persistent). Papavassiliou, Giorgos Flouris, Irini
Rename_Class(Existing, Persistent). Fundulaki, Dimitris Kotzinos, Vassilis
Rename_Class(Existing, Persistent). Christophides. On Detecting High-Level
Rename_Class(Existing, Persistent). Changes in RDF/S KBs. In Proceedings of
Del([Stuff subclass Existing]). the 8th International Semantic Web
Del([started_on domain Existing]). Conference (ISWC-09), to appear, 2009
Add([Persistent type Class]). Add([Stuff Vicky Papavassiliou, Giorgos Flouris,
subclass Persistent]). Add([started_on Irini Fundulaki, Dimitris Kotzinos,
domain Persistent]). 16/09/2009. Giorgos Vassilis Christophides. Formalizing
Flouris. 18. High-Level Change Detection for RDF/S KBs.
19Low-Level Versus High-Level Deltas. Technical Report TR-398, FORTH-ICS, 2009.
Purpose: A posteriori detect the 16/09/2009. Giorgos Flouris. 36.
differences (delta or diff) between
High-Level Change Detection in the Semantic Web.ppt
http://900igr.net/kartinka/anglijskij-jazyk/high-level-change-detection-in-the-semantic-web-63332.html
cсылка на страницу

High-Level Change Detection in the Semantic Web

другие презентации на тему «High-Level Change Detection in the Semantic Web»

«The animals» - SEA-HORSE. The animals which live in a SAVANNA. PENGUIN. WOMBAT. WHALE. FLAMINGO. BOBCAT. POLAR BEAR. REINDEER. SEAL. GIRAFFE. EMU. BEAR. KOALA. SNAKE. ELEPHANT. The animals which live in the polar regions. The animals which live in the desert. LIZARD. The animals which live in Australia. The animals which live in the OCEAN.

«Web 2.0» - Сообщество само вытесняет неактуальные материалы. «Желтофиоль». Огромная посещаемость и четкая сегментированность Целевой Аудитории – дорогая рекламная площадь. Примеры зарубежных проектов Web 2.0. Преимущества сервисов Web 2.0. На территории Рунета пока только появляются первые «ласточки». Примеры российских проектов Web 2.0.

«Технологии Web 2.0» - Применение технологий Web 2.0 сотрудниками департамента T&D. Может быть общедоступным или доступным группе. Технологии Web 2.0 для обучения. Корпоративное обучение 2009. Может быть платным или бесплатным. eLearning 1.0 eLearning 1.3 eLearning 2.0 http://elearningtech.blogspot.com/. Использование технологий web 2.0. в корпоративном обучении.

«Создание web-страниц» - Какова цель создания вашего сайта? Web - сайты и web - страницы. Где предполагается разместить сайт? Web - сайт. Общая характеристика данных, включаемых в сайт. Какие новые возможности предоставляет ваш сайт? <HEAD>. Текст. Рис. 1.1. Название сайта. <TITLE> </TITLE>. <BODY> </BODY>.

«The green movement» - Green color which is used by participants of movement as the general emblem, serves as a symbol of the nature, hope and updating. "Green" movement in the world. The main objective — to achieve the decision of global environmental problems, including by attraction to them of attention of the public and the authorities.

«Женщина the woman» - Значение понятия «женщина» в семье. Chicken’s mind- Куриные мозги. Пути пополнения лексической группы «женщина» в английском языке. Оценочная структура лексической единицы “женщина”. Пословицы. As great a pity to see a woman cry as a goose go barefoot. Женский интеллект. Человек = мужчина. « Der mann»- нем.

Тексты на английском

46 презентаций о текстах на английском
Урок

Английский язык

29 тем
Картинки