Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Towards a Semantic Wiki for Science

1,145 views

Published on

Ph.D. proposal presentation
Jacobs University
February 28, 2007

Published in: Technology, Education
  • Be the first to comment

Towards a Semantic Wiki for Science

  1. 1. Towards a Semantic Wiki for Science Ch. Lange Abstract Towards a Semantic Wiki for Science Wikis Research Proposal for a Ph.D. thesis Semantics Semantic Social Software Christoph Lange SWiM+ Prototype Jacobs University Bremen Knowledge (formerly International University Bremen) Representation Services February 28, 2007 Case Studies Conclusion 1 / 28
  2. 2. Abstract Towards a Semantic Wiki for Science Ch. Lange A semantic wiki for science — think “Wikipedia++” Abstract Represent scientific knowledge Wikis . . . which can be utilised by services Semantics . . . in a collaborative environment Semantic Social Software How to motivate users to contribute? SWiM+ Work plan towards the creation of SWiM+, the Prototype Semantic Wiki for Mathematics and Other Sciences: Knowledge Give the user services that exploit the knowledge Representation Access knowledge via “System Ontology” abstraction layer Services Case Studies Conclusion 2 / 28
  3. 3. Wikis in Science and Education Towards a Semantic Wiki for Science Ch. Lange Abstract Wikis Everything: e.g. Wikipedia Semantics Domain-specific: e.g. PlanetMath Semantic Social Anyone can contribute Software SWiM+ Quick and easy Prototype Create new knowledge . . . Knowledge Representation . . . or learn by browsing or searching Services Case Studies Conclusion 3 / 28
  4. 4. Wiki: Nice for Humans, but . . . Towards a Semantic Wiki for Science Ch. Lange Abstract Wikis Semantics Semantic Social Software SWiM+ Prototype Knowledge Representation Services Case Studies Imagine a scholar searching for: Conclusion sqrt{a^2 + b^2} = c, or: x^2 + y^2 = z^2 all theorems about triangles for which a proof exists 4 / 28
  5. 5. Semantic Markup Towards a Semantic Wiki for Science Ch. Lange Explicit markup of structures helps machines Abstract . . . e.g. intelligent search engines Wikis Tell the machine what you mean (= semantics) Semantics . . . not what it looks like! Semantic Social Markup languages available for: Software Mathematics: Content MathML, OpenMath, OMDoc, . . . SWiM+ Physics: PhysML Prototype Chemistry: CML Knowledge Representation Geodata: GML, SensorML, MarineML, ESML, . . . From top to bottom (roughly): less structures, more data Services Technology transfer: mathematics −→ other sciences Case Studies Conclusion 5 / 28
  6. 6. Semantic Markup Allows for Inference Towards a Semantic Wiki for Science Some dependencies among physical and mathematical theories Ch. Lange (think “course modules”): Abstract depends on discrete Wikis MP3 cosine audio Semantics transform Semantic Social Software SWiM+ Prototype Knowledge Representation Services Case Studies Conclusion 6 / 28
  7. 7. Semantic Markup Allows for Inference Towards a Semantic Wiki for Science Some dependencies among physical and mathematical theories Ch. Lange (think “course modules”): Abstract depends on discrete depends on Wikis MP3 orthogonal cosine audio matrix Semantics transform Semantic Social Software SWiM+ Prototype Knowledge Representation Services Case Studies Conclusion 7 / 28
  8. 8. Semantic Markup Allows for Inference Towards a Semantic Wiki for Science Some dependencies among physical and mathematical theories Ch. Lange (think “course modules”): Abstract depends on discrete depends on Wikis MP3 orthogonal cosine audio matrix Semantics transform Semantic Social Software depends on (transitive) SWiM+ Prototype (cf. ActiveMath [Melis et al., 2004]) Knowledge Inference gives you more output than input. Representation Services “I want to learn more about MP3 audio compression!” Case Studies “Fine, but do you know what an orthogonal matrix is?” Conclusion Description logic (DL) reasoners to this efficiently. 8 / 28
  9. 9. Semantic Markup: Who Wants to Write it? Towards a Semantic Wiki Pythagoras’ theorem again, in OMDoc/OpenMath: for Science <theorem id="pythagoras"> Ch. Lange <CMP>In a rectangled triangle, ...</CMP> <FMP> Abstract <OMOBJ> <OMA> Wikis <OMS cd="relation" name="eq"/> <OMA> Semantics <OMS cd="arith" name="plus"/> <OMA> Semantic <OMS cd="arith" name="power"/> Social <OMV name="a"/> Software <OMI>2</OMI> </OMA> SWiM+ <OMA> Prototype <OMS cd="arith" name="power"/> <OMV name="b"/> Knowledge <OMI>2</OMI> Representation </OMA> </OMA> Services <OMA> <OMS cd="arith" name="power"/> Case Studies <OMV name="c"/> <OMI>2</OMI> Conclusion </OMA> </OMA> </OMOBJ> </FMP> </theorem> 9 / 28
  10. 10. Semantic Markup: Who Wants to Write it? Towards a Semantic Wiki Pythagoras’ theorem again, in OMDoc/OpenMath: for Science <theorem id="pythagoras"> Ch. Lange <CMP>In a rectangled triangle, ...</CMP> <FMP> Abstract <OMOBJ> <OMA> Wikis <OMS cd="relation" name="eq"/> <OMA> Semantics <OMS cd="arith" name="plus"/> <OMA> Semantic <OMS cd="arith" name="power"/> Social <OMV name="a"/> Software <OMI>2</OMI> Theorem (pythagoras) </OMA> SWiM+ <OMA> In a rectangled triangle, . . . <OMS cd="arith" name="power"/> Prototype <OMV name="b"/> a2 + b 2 = c 2 Knowledge <OMI>2</OMI> Representation </OMA> </OMA> Services <OMA> <OMS cd="arith" name="power"/> Case Studies <OMV name="c"/> <OMI>2</OMI> Conclusion </OMA> </OMA> </OMOBJ> </FMP> </theorem> 10 / 28
  11. 11. The Semantic Author’s Dilemma Towards a Semantic Wiki for Science Ch. Lange Abstract Wikis macro-perspective: “OK, if I make the effort, others can find Semantics my formulæ using Google++” Semantic Social micro-perspective: “. . . but what do I gain from it?” Software SWiM+ Give the author added value in his particular situation! Prototype [A. Kohlhase, Müller 2007] Knowledge Representation Services Case Studies Conclusion 11 / 28
  12. 12. Semantic Social Software Towards a Semantic Wiki for Science Ch. Lange Abstract Wikis Semantics Semantic Social Software SWiM+ Prototype Knowledge Representation Services Case Studies Conclusion [Spivack 2003, Schaffert 2006] Here: Semantic Wiki + OMDoc and relatives Integrate services, share their added value with the user 12 / 28
  13. 13. Applications of SWiM+ : Stakeholder Analysis Towards a Semantic Wiki for Science Ch. Lange SWiM+ Abstract • Semantic Markup Wikis • Services Semantics Semantic Social Software SWiM+ Prototype Scientists Scholars Semantic Web Knowledge • Research • Learn • Access Representation • Collaboration • Search/Find • Reuse Services Case Studies Conclusion 13 / 28
  14. 14. Applications of SWiM+ : Science Towards a Semantic Wiki Scientists want to formalise their ideas and develop theories for Science collaboratively. Ch. Lange Example Abstract Wikis E. In. Stein wants to elaborate on his hypotheses about Semantics “relativity” Semantic Social Software . . . and decides to build on the theory of “gravitation” SWiM+ . . . developed by his colleague N. Ew. Ton Prototype . . . which is still under development, though! Knowledge Representation If Ton changes some basic assumptions on gravitation, Services . . . Stein’s considerations might become invalid. Case Studies Conclusion SWiM+ must be able to manage dependencies! (cf. [Müller 2006]) 14 / 28
  15. 15. Applications of SWiM+ : Education Towards a Semantic Wiki for Science Ch. Lange Adapt to the previous knowledge of a learner Abstract Needs to find what he wants by browsing or searching Wikis Semantics An example from algebra (each theory on one wiki page): Semantic Social I already know the Show me the di- Software group concepts depended rect dependencies! ring (group and monoid) upon SWiM+ Prototype monoid Knowledge Representation Show me all depen- I want to know, but semigroup dencies! (semigroup, remind me later Services too) Case Studies Conclusion 15 / 28
  16. 16. Applications of SWiM+ : Semantic Web Towards a Semantic Wiki for Science Ch. Lange Weave the Semantic Web with SWiM+ — make scientific Abstract knowledge reusable on the web Wikis Semantics Export knowledge extracted from markup to standard Semantic formats (RDF) Social Software . . . which can be imported into other Semantic Web applications SWiM+ Prototype Integrate external services Knowledge e.g. MathWebSearch, our formula search engine Representation [M. Kohlhase, Șucan 2006] Services Case Studies Conclusion 16 / 28
  17. 17. SWiM (SWaM?) Prototype = IkeWiki + OMDoc Towards a Semantic Wiki First experiment (my diploma thesis) for Science Ch. Lange editing, presentation, navigation; partial system ontology Abstract Wikis Semantics Semantic Social Software SWiM+ Prototype Knowledge Representation Services Case Studies Conclusion 17 / 28
  18. 18. Scientific Knowledge Representation Towards a Semantic Wiki I assume three levels of scientific knowledge [M. Kohlhase, for Science 2006]: Ch. Lange Objects symbols, numbers, equations, molecules, . . . Abstract Statements axioms, hypotheses, measurement results, Wikis examples; relationships: “proves”, “defines”, Semantics “exemplifies”, . . . Semantic Social Theories collections of interrelated statements, determine Software context: “What does the symbol h mean?” SWiM+ Prototype Successfully applied to mathematics Knowledge Representation OMDoc (M. Kohlhase) Services Transferred to physics with small additions Case Studies PhysML (Stamerjohanns et al.) Conclusion Anticipation: also valid for chemistry (Murray-Rust), geosciences (Baumann, Unnithan), . . . 18 / 28
  19. 19. System Ontology Towards a Semantic Wiki Ontology: formal, explicit specification of a for Science conceptualisation (of a specific domain) Ch. Lange System ontology (here): ontology of the systems built Abstract around a markup language (meta level!) Wikis Defines OMDoc Semantics how to represent knowledge in Models Maths Semantic Social Software Theory SWiM+ lives in imports OWL-DL Prototype Statement is a is a implementation, Knowledge Representation is a with generic relation Symbol is a Proof Services defined by is a proves types dependency Case Studies and containment Definition Assertion Conclusion (both transitive) exemplifies Example exemplifies exemplifies 19 / 28
  20. 20. Where is the Knowledge? Towards a Semantic Wiki Semantic wiki: One page = one concept for Science SWiM+ : One page = one statement or theory Ch. Lange (small, reusable pages, but not too small) Abstract Semantic Web tools need more explicit representation of Wikis the knowledge (extract from markup!) Semantics Semantic Example Social Software A wiki page: Extracted RDF triples (= graph): SWiM+ proves Prototype <omdoc> Proof Theorem <proof id="pyth-proof" type type Knowledge for="pythagoras"> Representation proves ... pyth-proof pythagoras Services </proof> Case Studies </omdoc> <pyth-proof, rdf:type, omdoc:Proof> Conclusion <pyth-proof, omdoc:proves, pythagoras> (omdoc:* → OMDoc system ontology) 20 / 28
  21. 21. Service Programming Interface: Access the Knowledge Towards a Semantic Wiki for Science Ch. Lange Make generic services work for all markup languages (= all scientific domains) Abstract “SciML” researchers will identify common traits of the Wikis system ontologies (expected: theory/statement/object, Semantics containment, dependency) Semantic Social Formalise that upper system ontology for SWiM+ ! Software Analyse services to design programming interface: SWiM+ What knowledge do they need? Inference required? Prototype Knowledge Also model a system ontology of the wiki; Representation inference over both Services Many user questions to a proof page ⇒ must be hard to Case Studies understand ⇒ auto-create examples for special cases Conclusion 21 / 28
  22. 22. Which Services? Added Value? Towards a Semantic Wiki for Science Ch. Lange Plan services using added-value analysis Abstract [A. Kohlhase/Müller 2007]: Wikis 1 Core problem Semantics 2 Core solution Semantic Social 3 Benefits/sacrifices Software 4 . . . may trigger new core problems SWiM+ Their objective: Motivate user to take action! Prototype Provide benefits while the user contributes Knowledge Representation In semantic wikis also: instant gratification Services “Thanks for your contribution, and look . . . ” Case Studies Conclusion 22 / 28
  23. 23. Example: Auto-Completion Service Towards a Semantic Wiki for Science Support easy linking (wiki ) Ch. Lange facilitate editing text (markup ) Abstract Suggest links to appropriate targets Wikis Semantics <proof for="p_ Semantic Social Software SWiM+ Prototype Knowledge Representation Services Case Studies Conclusion 23 / 28
  24. 24. Example: Auto-Completion Service Towards a Semantic Wiki for Science Support easy linking (wiki ) Ch. Lange facilitate editing text (markup ) Abstract Suggest links to appropriate targets Wikis Semantics <proof for="p_ Semantic partial-diff-eqn Social proton Software pythagoras SWiM+ Prototype Knowledge Representation Services Case Studies Conclusion 24 / 28
  25. 25. Example: Auto-Completion Service Towards a Semantic Wiki for Science Support easy linking (wiki ) Ch. Lange facilitate editing text (markup ) Abstract Suggest links to appropriate targets Wikis Semantics <proof for="p_ Semantic partial-diff-eqn Social proton Software pythagoras SWiM+ Prototype Find all known instances of the range of the relation Knowledge Representation Proof–proves–? starting with “p” Services First service to be implemented, because: Case Studies does not require DL reasoning (no transitivity) Conclusion easy to evaluate productivity of volunteers 25 / 28
  26. 26. More Ideas for Services Towards a Semantic Wiki for Science Ch. Lange Already specified: Abstract Dependency Management: simple assistance Wikis (later: locutor [Müller 2006]) Semantics Dependency Graph Navigation (see above) Semantic Social Software “Edit-in-place” (section-wise editing of long documents) SWiM+ Planned: Prototype QMath, an easy syntax for formulae Knowledge Representation MathWebSearch interface Services Added-value analysis will give more ideas. Case Studies Conclusion 26 / 28
  27. 27. Case Studies Towards a Semantic Wiki for Science Ch. Lange Abstract Science Our group is working on the cross-domain Wikis integration of scientific markup languages Semantics test SWiM+ in a heterogeneous environment Semantic Social Education M. Kohlhase’s slides for “General Computer Software Science” available as s EX (OMDoc-like T SWiM+ A LTEX) Prototype convert, import, evaluate learning and Knowledge Representation interaction services Services Case Studies Conclusion 27 / 28
  28. 28. Better Wikipedia, Better Semantic Web Towards a Semantic Wiki for Science benefit n ve Ch. Lange -e ak e Abstract br Wikis Semantics SWiM+ Formal Methods Semantic (2009) Social Software my challenge SWiM+ Prototype Wikipedia Semantic Web Knowledge Representation (today) WWW Services (Web 1.0) Case Studies Conclusion sacrifice 28 / 28

×