Bridging Semantic and Legacy Web Exploration: Bridging Semantic and Legacy Web Exploration: The Reasoning and Scalability Point of View The Reasoning and Scalability Point of View

blaredsnottyΤεχνίτη Νοημοσύνη και Ρομποτική

15 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

97 εμφανίσεις

1
1
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
31
31
.
.
8
8
.20
.20
1
1
0
0
WIRSS /Web
WIRSS /Web


KR3 2010, Toronto, Canada
KR3 2010, Toronto, Canada
Bridging Semantic and Legacy Web Exploration:
Bridging Semantic and Legacy Web Exploration:
The Reasoning and Scalability Point of View
The Reasoning and Scalability Point of View
Michal
Michal
Tvaro
Tvaro
ž
ž
ek,
ek,
M
M
á
á
ria Bielikov
ria Bielikov
á
á
Slovak University of Technology
Slovak University of Technology
in
in
Bratislava
Bratislava
2
2
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Presentation outline
Presentation outline
•Motivation and goals
•What are we trying to do?
–Legacy and Semantic Web exploration
–Web information retrieval support systems
•The things we have trouble with
–Query scalability, changing information
•So what now? [room for your ideas]
3
3
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
The Semantic Web today
The Semantic Web today
•Semantically annotated data
•Search engines
–Sindice.com
•Query builders
–OpenLink
•Basic browsing
–Tabulator
–ZitgistViewer
taken from linkeddata.org
4
4
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Motivation
Motivation
•End‐user problems with the search process
–Query construction and refinement
(keyword guessing, limited refinement support)
–Results browsing and selection
(short snippets, dead links, out‐of‐date data)
–Result exploration and understanding
(big documents, no relations, context loss)
–Orientation and resource revisitation
…in both the Semantic and Legacy Web
5
5
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
•Adaptive Web (1996, 2001, 2007)
•Semantic Web (2001, 2006)
•Exploratory search (2006)
•Social Web (2005)
•Browser/search engine support for
–Searching
–Navigation
–Visualization
The broader web perspective
The broader web perspective
6
6
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Goals
Goals
•Enable Semantic Web search and exploration
•Merge Legacy and Semantic Web experience
•Support exploratory search features
…for end‐users not just professionals…
…but does it scale (at all)??
7
7
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Shifting the browser paradigm
Shifting the browser paradigm
•Faceted personalized browsing(2005, 2007)
–Delegate querying, indexing, crawling, reasoning 
services to third‐party providers (Google, Sindice)
•Ontological information representation (OWL)
•Integrated support for
–Search (query construction & modification)
–Navigation (guidance, annotation)
–Exploration (content adaptation and visualization)
8
8
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
•Complex query construction
Visual query construction
•Limited end‐user exploration
Multi‐paradigm exploration
•Information overload
Personalized recommendation
•No default visualization
Adaptive view generation


NextGen
NextGen


Semantic Web Browser
Semantic Web Browser
9
9
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Facet adaptation & recommendation
Facet adaptation & recommendation
10
10
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Collaborative resource annotation
Collaborative resource annotation
11
11
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Search history trees
Search history trees
12
12
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Semantic history maps
Semantic history maps
13
13
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Incremental graph exploration
Incremental graph exploration
14
14
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
What about legacy data?
What about legacy data?
•Each page is considered a SW resource
–Pre‐indexed metadata via search services
•Usage metadata
–Tracked and provided by external proxy servers
(Kramár, Barla, Bieliková, 2010)
–Whovisited what and when
15
15
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Lightweight semantics discovery
Lightweight semantics discovery
•Content metadata
–Taken from existing CMS (e.g., if in RDF or OWL)
–Extracted from existing pages and/or sites
•The model:
–Core web page content resources
–Navigation menus classifications
–Contextual links annotations
16
16
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Adding semantics to legacy content
Adding semantics to legacy content
1.Metadata extraction
–Term extraction, disambiguation and annotation
–Cross‐reference via public bookmarking systems
2.Hierarchical classification
–Hierarchical web site structure extraction
–Term matching to external taxonomies 
(delicious)
3.Annotation extraction
–Incoming contextual link extraction, term 
extraction, matching and annotation
17
17
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Client focused architecture
Client focused architecture
WCF
View
renderers
View
renderers
View
renderers
Event
logs
User
model
Event
log
SemanticLog
Logging
service
Steltecia
Repository
access service
SPARQL
SeRQL
User
models
Domain
models
Factic
Faceted search
engine
Support
Services
Event
handlers
Personalization
engine
Support
plugins
Service
adapters
18
18
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
The main issues were
The main issues were


•(Meta‐)Data representation (logs, domain, 
user model, relevance model, access control)
•Network delays (client‐server, server‐db)
•Repository scalability (with respect to hw)
•Querying scalability, QL expressiveness
•Reasoning capability and scalability
19
19
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Data representation
Data representation
•Log data –stream processing?
•Access control –mapping to domain/user 
data
•User model –mapping to domain model
•Domain model
–Interlinking with / extension of other models
–Static moment in time vs. evolving changes
(temporal databases? querying?)
–Geographical information (spatial queries)
•Social networks and communities (partitions?)
20
20
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Querying issues
Querying issues
•Sesame 1.2 vs. 2.1 (and Java)
•SeRQLvsSPARQL
–What about updates?
–What about aggregation operators, e.g. ‘COUNT()’
•Queries take too long
–Too complex (emulate lack of reasoning)
–Too many results (emulate aggregations)
–Too complex results (too much data in general)
21
21
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Reasoning issues
Reasoning issues
•Is it worth it? [owlimquery improvements]
•Offline vs. online reasoning
•How about changes?
–Propagate to the server side
–Propagate to the client side
•Discrete vs. continuous values
22
22
Bridging Semantic and Legacy Web Exploration
Bridging Semantic and Legacy Web Exploration
Conclusions
Conclusions
•Adaptive Social Semantic Web Exploration
–Interactive multi‐paradigm exploratory search
–Personalized recommendation for individual users
–Visual query construction
–Adaptive view generation
•Need for scalable back‐end services