WP8: User Centred Applications
Enrico Motta, Marta Sabou, Vanessa Lopez, Laurian Gridinoc, Lucia Specia
Knowledge Media InstituteThe Open UniversityMilton Keynes, UK
WP8 Goals and Tasks
• Objective:– To provide and evaluate concrete applications of OK to
support user tasks on the Web, such as knowledge retrieval and ontology-assisted browsing.
• Tasks:– T8.1. Semantic Browsing
• Evolve Magpie so that it does not rely on design time ontology selection
– T8.2. Ontology Based Query Answering• Evolve AquaLog towards domain independent QA
– Evaluating the value of OK Technology• Compare standard and OK-enabled versions of both systems
Outline
• Vision:– “Open” is core to novel Semantic Web applications– Novel technical challenges arise
• Building novel applications within OpenKnowledge:– New methods:
• Dynamic ontology mapping
– Providing more semantic data:• Folksonomy enrichment
The SW gets BIGGER
Lee, J., Goodwin, R. (2004) The Semantic Webscape: a View of the Semantic Web. IBM Research Report.
The Semantic Web registered a 300% growth in 2004 alone, thus outpacing the growth of the Web itself.
Access Gateways exist
Example1: AquaLog
1. NL Question
2. Linguistic interpretation
3. Ontology based interpretation
4. Answer
Limited to the domain and data provided by a single ontology
Example1: AquaLog
Cross domain QA: Selects and combines relevant information from multiple ontologies:• automatically locate ontologies • map user terminology to ontologies• integrate info from different ontologies (mapping)
PowerAqua: QA on the 'open' Semantic Web
Example2: Magpie
NL QuestionOntology concepts
Instances highlighted according to their type
Example2: Magpie
Limited to the domain and data provided by a single ontology
PowerMagpie: Semantic browsing on the 'open' SW
Open semantic browsing: Dynamically selects and combines relevant information from multiple ontologies:• automatically locate ontologies • integrate info from different ontologies (mapping)
New Tools are OPEN
• … with respect to the topic domain– Instead of deciding the domain at design time– Let the user decide the domain of interest at run-time– Thus: Lower the cost of user participation
• … with respect to the explored data– Instead of “hard-wiring” one knowledge sources at design
time - smart databases– Dynamically select and make use of multiple,
heterogeneous knowledge sources:• Online available ontologies/semantic data• Non-semantic data, e.g., folksonomies
– Thus: Lower the cost of data integration
Key Paradigm Shift
Invited talks and papers:
Motta, E., Sabou, M. "Next Generation Semantic Web Applications". ASWC’06. Motta, E., Sabou, M. "Language Technologies and the Evolution of the Semantic Web". LREC’06
Source of Intelligence:
• Early Semantic Web tools:• A function of sophisticated, task-centric problem solving
• New Tools: •A side-effect of size and heterogeneity (Collective Intelligence)
What is needed?
Dynamic Ontology Selection
Ontology Modularization
Dynamic Ontology Mapping
Current work focuses on
user-mediated ontology selection
Current work assumes user involvement
Current work:• design-time mapping of complete ontologies• assumptions on the domain and structure of the ontologies
Outline
• Vision:– “Open” is core to novel Semantic Web applications– Novel technical challenges arise
• Building novel applications within OpenKnowledge:– New methods:
• Dynamic ontology matching
– Providing more semantic data:• Folksonomy enrichment
Achievements – at a glance
• Ontology Matching– Two dynamic ontology matching algorithms
• Run-time matching of knowledge structures• No assumptions on domain, structure etc.
– Core to our tools and to the OK infrastructure– Defined, implemented, documented, partially tested
• PowerMap – part of PowerAqua• MatchMiner - matching by using the Semantic Web as
background knowledge
• Acquiring semantic data– A Hybrid Algorithm for learning relations from text– Semantic enrichment of folksonomies by exploring online
ontologies
PowerMap: core of PowerAqua
Keywords
OntologyTriples
1. Ontology identification•Syntactic mapping
2. Extracting (clusters of) triples•Semantic mapping
3. Filtering triples
PowerMap
–Lopez, V., Sabou, M., Motta, E. "Mapping the Real Semantic Web on the Fly". ISWC’06.–Reported in deliverables D3.1. and D4.1.
• rely on online ontologies (Semantic Web) to derive mappings• ontologies are dynamically discovered and combined• does not require any a priori knowledge about the domain• returns semantic relations as mappings
A Brel
Semantic Web
MatchMiner
•M. Sabou, M. d’Aquin, E. Motta, “Using the Semantic Web as Background Knowledge in OntologyMapping", Ontology Mapping Workshop, ISWC’06. – Best Paper Award•Reported in Deliverable D4.1.
Evaluation: 1600 mappings, two teamsAverage precision: 70% (comparable/better than standard)
(derived from 180 different ontologies)
Matching AGROVOC (16k terms) and NALT(41k terms)
Large Scale Evaluation
M. Sabou, M. d’Aquin, W.R. van Hage, E. Motta, “Improving Ontology Matching by Dynamically Exploring Online Knowledge", submitted for review, 2007.
Semantic Folksonomy Enrichment
Tags
{camera, digital, photograph} {damage, flooding, hurricane, katrina, Louisiana} Clusters
digital
cameraphotographtakenWith
Ontology
NLP/Clustering
Find and combine Online ontologies
L.Specia, E. Motta, "Integrating Folksonomies with the Semantic Web", submitted for review, 2007.
Examples
Examples
Summary
• The growing SW allows opening up applications– With respect to their domain– And the exploited data sources
• Novel (dynamic) methods are required for:– Ontology selection, matching and modularization
• Dynamic and approximate ontology matching:– Is core to both our applications and the OK framework– We provided two novel algorithms for this topic
• Folksonomy enrichment– Is a way to get more data for our tools– We provided an algorithm based on ontology matching
Next Steps
• Finalize the prototypes:– PowerAqua (M18)
• Integrate PowerMap within PowerAqua• Make use of the semantically enriched folksonomies
– Semantic Browser (M18)• Combine ontology selection, matching and modularization
techniques
• Evaluate our applications (M24, M36):– When based on mainstream SW technology– Extended to take advantage of the OK infrastructure
Thank you!
Vision Papers
• Motta, E., Sabou, M. (2006). "Next Generation Semantic Web Applications". ASWC.
• Motta, E., Sabou, M. (2006). "Language Technologies and the Evolution of the Semantic Web". LREC 2006
• Motta, E. (2006). "Knowledge Publishing and Access on the Semantic Web: A Socio-Technological Analysis". IEEE Intelligent Systems, Vol.21, 3, (88-90).
• V. Lopez, E. Motta and V. Uren (2006) “PowerAqua: Fishing the Semantic Web”, ESWC’06.
Ontology Matching
• Lopez, V., Sabou, M., Motta, E. (2006). "Mapping the real semantic web on the fly". ISWC.
• Sabou, M., D'Aquin, M., Motta, E. (2006). "Using the semantic web as background knowledge for ontology mapping". ISWC 2006 Workshop on Ontology Mapping.
• M. Sabou, M. d’Aquin, W.R. van Hage (2007), E. Motta, “Improving Ontology Matching by Dynamically Exploring Online Knowledge", submitted for review.
Relation Learning/Folksonomy Enrichment
• L. Specia, E. Motta (2006): “A hybrid approach for relation extraction aimed to semantic annotations”. 7th Flexible Query Answering Systems (FQAS).
• L. Specia, E. Motta (2006): “A hybrid approach for extracting semantic relations from texts”. Workshop on Ontology Learning and Population (OLP2)
• L.Specia, E. Motta (2007), "Integrating Folksonomies with the Semantic Web", submitted for review, 2007
Related NeOn papers
• Ontology Selection– Sabou, M., Lopez, V., Motta, E. (2006). "Ontology
Selection for the Real Semantic Web: How to Cover the Queen’s Birthday Dinner?". EKAW 2006
• Ontology Modularization– D'Aquin, M., Sabou, M., Motta, E. (2006).
"Modularization: A key for the dynamic selection of relevant knowledge components". ISWC 2006 Workshop on Ontology Modularization