Extended abstract
General goal, approach and guideline. The general aim of ScalIntegr, the long-term project that I propose, is to investigate and develop models, techniques and tools that support scalable manual and automatic ways of retrieving, extracting, representing, integrating, sharing, comparing, and managing information from various sources (documents, knowledge bases, knowledge providers), in a precise, normalized and semantically well interconnected manner. To do so, this project should refine my work aimed to support the collaborative building of shared KBs approach and extend it to integrate more distributed and/or automatic approaches to knowledge creation. However, for scalability purposes, my guideline for all the research tasks of this project will still be the same: trying to reduce the implicit redundancies or inconsistencies between the knowledge objects (e.g., categories or statements) and, more generally, increase the semantic organization of these objects.
Themes of research. This project will focus on finding and supporting ways to guide or automatize
Collaborations and applications. This project, which is a broadening of my research work since 2000, fits well in the research programs of both the KEWI (Knowledge Engineering and Web Intelligence) team of the I3S laboratory and the IC3 (Ingénierie de la Connaissance, de la Cognition et de la Coopération) team of the IRIT laboratory. It would also support shared research and tools with related teams at the I3S laboratory (in particular RAINBOW but also RL, OASIS, AOSTE, Bioinfo and MC3) and IRIT laboratory (SMAC, SIG, ILPL, LiLaC and RPDMP). Other possible applications will be in software engineering and bioinformatics, and will involve semantic grids and peer-to-peer networks. Most applications will be related to knowledge base sharing, the Semantic Web and corporate memories.
Table of Contents
The general goals of this project are shared by many research projects, including major ones such as Knowledge Web and NeOn (Networked Ontologies) which are funded by the European Commission to make the Semantic Web vision a reality. However, these projects mainly focus on techniques aimed to support a "semi-independent knowledge representation approach". Indeed, most projects prefer to focus on exploiting already existing resources and adopt the current main paradigm and languages advocated by the W3C for the Semantic Web (Berners-Lee et al., 2001) (Shadbolt et al., 2006), even though many of their basic flaws are well discussed, as in (Kalfoglou et al., 2004) and (Patel-Schneider, 2005). It can also be argued that, at least nowadays, the semi-independent approach is socially necessary for scalable knowledge sharing and retrieval. Furthermore, genuinely supporting the "cooperative building of large knowledge bases" (or in other words, the "shared KB approach") is an ambitious long term goal that requires the design and implementation of many techniques. At a general level, the originality of ScalIntegr is that it will (i) extend my work on the shared KB approach since it is technically necessary for scalable knowledge sharing and retrieval, and (ii) extend it to better integrate the semi-independent approach. Extensions to semi-automatic knowledge extraction and merging techniques will also be made; such extensions are needed for both approaches. The particular approaches proposed below are also original. Each of the tasks that I propose should be done in collaboration with the team members of KEWI or IC3.
This project investigates techniques to support an efficient avenue for the incremental building of a "Knowledge Web" or genuine "Semantic Web" (on the Internet, within an intranet, or in a peer-to-peer network), that is, a well organized semantic network permitting objects (e.g., conceptual categories, beliefs, definitions, ideas, tasks, data structures, persons, organizations and products) to be accessed and compared precisely and efficiently. Each of such networks may be distributed, that is, it may span several KBs but ideally these KBs should be well interconnected and hence should not have implicit redundancies or inconsistencies. This is why I refer to each of such networks, as one global "virtual" well-organized KB.
The current nonexistence of such a network is the cause of the well acknowledged problem of information overload or difficulty of finding precise and relevant information in intranets or the Internet. For example, questions such as "What are the arguments and objections for the use of an XML-based format for the exchange of knowledge representations?", "What are all the tasks that should be done in software engineering according to the various existing 'traditional system development life cycle' models?" and "What are the characteristics of the various theories and implemented parsers related to Functional Dependency Grammar and how do these theories and parsers respectively compare to each other?" cannot be adequately answered by giving a list of documents or an unorganized list of all sentences related to these questions. They require presenting and allowing the browsing of a semantic network: (i) for the first question, a network with argumentation, objection and specialization relations, (ii) for the second question, a subtask hierarchy of all the advised tasks, and (iii) for the third question, a network with specialization relations between the various objects or attributes related to the theories and parsers.
My research report (the "report on past research works" that is included in this application) introduced a basis for the dream of Buckingham et al. (2007) - creating an Internet based infrastructure supporting a more effective dissemination, debate, and analysis of ideas than the current system of article publication - and the related dream of Hillis (2004) - a "Knowledge Web" to which teachers and researchers could add "isolated ideas" and "single explanations" at the right place, with "mechanisms for credit assignment, usage tracking, and annotation that the Web lacks" thus supporting a much better re-use and evaluation of the work of a researcher than the current system of article publishing and reviewing. To support these dreams and the closely related goal of cooperatively-built virtual well-organized KBs, research should be undertaken on various points related to knowledge representation, organization, presentation, extraction and integration. For all of them, the general guideline of ScalIntegr is that ideally there should be no implicit redundancies between objects. Some alternative formulations are: (i) the objects should be represented as precisely and uniformly as possible, (ii) they should be as small and explicitly interconnected by semantic relations as possible, at least identity and specialization relations, and (iii) the bigger and more organized the KBs, the easier it is for a software to align/merge these KBs or guide the users in entering precise and re-usable knowledge. These formulations can be seen as my work hypothesis. The designed methods and tools will be tested and refined in the context of applications such as the cooperative building of various semantically-well-organized states of the art. The implementations will re-use an integration of the WebKB-2 server (Martin & Eklund, 2001) with the tools of the KEWI or IC3 teams.
There are many automatic techniques and tools to extract knowledge from texts/databases or align conceptual categories from different KBs. They are useful but insufficient for precision-oriented knowledge retrieval and sharing. For example, Euzenat & Shvaiko (2007) give an evaluation of ontology alignment techniques and conclude that these tools are quite understandably imperfect although sufficient for certain applications. They recognize the need for the approach I advocate: the design of (semi-)formal KBs letting people and software agents directly exploit and save new knowledge or object alignments, that is, query, complement, annotate and evaluate the existing objects, guided by these large and well-organized KBs.
To support scalable ways of sharing knowledge between KBs (and hence the creation of a global virtual well-organized KB), the general guideline has two implications. A first one is that current knowledge extraction and alignment techniques should be combined and refined according to methodological and normalization guidelines (those I gathered so far plus new ones). This task, which is described in Section 5, will be guided by the already integrated ontologies in the destination KB. Given the guidelines, each integration of each ontology should not lead to a loss of information; this aspect will permit to further automatise the integration of newer versions of that ontology and the re-application of corrections made to previous versions of that ontology when such corrections are still necessary.
A second implication is that it should not matter which (non-virtual) KB an intranet/Internet user or agent chooses to query or update first. Hence, a research question is "how to support the replication of knowledge between KBs, whether they are publicly accessible, within intranets or within a peer-to-peer network?". More precisely, two constraints are: 1) object additions/updates made in one KB should be replicated into all the other KBs that have a scope (domain of interest) which covers the new objects, and 2) a query for which the content of a KB will not yield a complete answer (with respect to the content of the virtual global KB) should be forwarded to the appropriate KBs. Point 1 implies that two KBs having shared sub-scopes should have the same objects for these scopes. The second point implies that, in each KB, when an object within the scope of a KB is related to an object outside the scope of the KB, a reference to this last object in another KB should be used.
To satisfy those two constraints, ScalIntegr will start with an approach described in (Martin & Eboueya, 2008) and summarized in Table 1. More precisely, I propose to refine and implement this approach in two contexts: (i) for KB servers on the Web, and (ii) for KB servers in peer-to-peer networks (e.g., for a semantic grid) with the assumption that each user has its own KB server. The main difference is that a peer-to-peer network will permit to implement systematic push/pull mechanisms instead of relying on KB servers to regularly check the KBs of other servers and integrate new additions. This also means that in a peer-to-peer network intra-KB collaboration protocols can also be integrated in inter-KB collaboration protocols. To do so, the current protocols of WebKB-2 (Martin et al., 2005) should first be combined with those of Co4 (Euzenat, 1996). I'll undertake this research in collaboration with the members of the KEWI team working on the research theme "knowledge sharing and exchange". The work related to peer-to-peer networks will be done in collaboration with A.Pr. Eboueya (L3I, University of La Rochelle) with whom I already collaborate (as shown by my list of publications) and who uses semantic grids for e-learning purposes.
The approach of Table 1 seems the simplest approach. Indeed,
(i) the approaches used in distributed databases would not work since KBs do not have any
fixed conceptual schema (they are composed of large, explicit and dynamically modifiable
conceptual schemas), and
(ii) a fine-grained classification (via an ontology) of all the objects is necessary
for precision-oriented information retrieval: indexing KB servers according to topics or domains is a far too coarse indexation to retrieve knowledge from distributed servers (e.g., knowledge about "neurons" or "hands"
can be relevant to many domains).
To the best of my knowledge there is no other project addressing the above two constraints.
The works dealing with "Ontology Evolution in Collaborative Environments", e.g.,
(Vrandecic et al., 2005) and (Noy & al, 2006) - or (Rousset, 2004) and
(De Roure et al., 2005) in a peer-to-peer or semantic grid context -
are solely about accepting/rejecting and integrating changes made in other KBs, they are
not about making these KBs have an equivalent content for their shared sub-scopes.
|
The scope of each server is defined by a reference collection of objects
(conceptual categories or statements)
and an optional maximum depth for the chain of relations from these objects
(for a totally general server, this collection is reduced to "Thing",
the most general conceptual category imaginable).
Each server periodically checks more general servers, competing
servers and slightly more specialized servers, and 1) integrates (and hence mirrors) all the objects generalizing the objects defined in the reference collection, 2) integrates all the objects (and associated direct relations) more specialized than those in the reference collection until it reaches a maximum specialization depth if one has been specified (if so, the URL of the object is stored instead of the object), 3) also stores the URLs of the direct specializations of the generalizations of the objects in the reference collection (this is needed for any object in the global virtual network to be directly or indirectly referred to). |
My research report quickly presented my creation/integration and enhancement of (i) knowledge representation methodologies, best practices and normalization procedures (Martin, 1993, 1996, 2000), (ii) ontologies (Martin, 2003) (Martin & Eboueya, 2007), and (iii) knowledge-based collaboration and valuation protocols (Martin et al., 2005). It also presented their integration into expressive but high-level languages (For-Links, Frame Conceptual Graphs and Formalized English) (Martin, 2002) or into the shared ontology, as with my representation of KADS models and my additions to them for explanatory information (Martin, 1996). The refinement of my previous research on these points is necessary for supporting the creation and extension of well organized semantic networks.
These points fit perfectly in some research themes of both KEWI and IC3: the themes "Models of knowledge and texts" and "Models and plans" of IC3, and the themes "Knowledge modelling and inferencing" and "knowledge sharing and exchange" of KEWI. The rest of this section shows this with respect to the goals of the last theme: the abstraction of mechanisms shared by different Semantic Web languages, the definition of a generic core language allowing extensions from all languages based on it, the design of a generic graph-based knowledge representation platform, the design of KB validation mechanisms based on semantic constraints, and the optimal organization of ontologies.
Fully and correctly importing or exporting between these notations and other languages such as KIF, CL and W3C languages such as RDF+OWL and N3 is a difficult but important research work since it means translating them from/to languages with different structures (lower-level structures or non-graph based structures) and different expressiveness. Yao & Etzkorna (2006), or Dieng-Kuntz & Corby (2005) of the INRIA Edelweiss team with which the KEWI team regularly collaborate, have already proposed methods to convert knowledge between the RDF model and its related subset within the Conceptual Graph model. Taking more expressiveness into account implies making more ontological choices and exploiting as well as complementing the ontology on which the source and destination knowledge representations (KRs) are based. Thus, this research involves ontology translation (Corcho, 2005) (Euzenat & Shvaiko, 2007) and hence overlaps with the one on ontology integration. This research wil be done in collaboration with the EXMO team (led by Euzenat and including Corcho) and will be guided by methodological and normalization guidelines and protocols. To come up with them and, more generally, to guide knowledge entering, the current guidelines and protocols of WebKB-2 will be refined and extended. For example:
Formalized English (FE) (Martin, 2002) is currently only a syntactic variant of FCG and hence does not lead to translation research problems. The downside is that this FE is not as easy to read or write as other formal or informal controlled languages which are often proposed to ease knowledge representation. A first necessary extension is, while keeping FE formal, to allow it to be more English-like by extending its grammar and exploiting the ontology. Then, via a set of options, each user should be allowed to see or use more and more informal extensions to FE. Ultimately, the use of simple but common English sentences should be allowed for knowledge entering and querying, with their translations into a formal version of FE being shown to the user for her to check the interpretation being made of her sentence or select between different interpretations. This option-based approach will offer the advantages of all "controlled languages", formal and informal. A last option will allow the parsing of several sentences without the user having to check the interpretations; in this case, a semi-formal version of each sentence will be stored. This means that the data model and the graph-matching operators will have to be updated to accept more informal elements. Words (instead of category identifiers) are already accepted by WebKB-2 but, for scalability and efficiency purposes, the data model needs to be made more generic (e.g., my recent investigations led to the conclusion that words and category identifiers should not be stored into different data structures and should be organized into a same specialization hierarchy - unlike in WordNet for example). This research overlaps with the one on knowledge extraction from natural language described in Section 5.
The results of my research ease the collaborative or personal representation and evaluation of formal and informal information (e.g., argumentation structures and information about users, including their contributions) in a precise and uniform way. The represented information should then be usable for knowledge querying, filtering and presentation purposes. WebKB-2 already proposes many presentation options and permits the specification of filtering constraints using FL or FCG. For example, during browsing or in the results of queries, certain categories or statements created by authors having certain characteristics are not displayed or are displayed in small fonts. However, the filtering specifications must be manually made, and no set of commands nor conceptual categories are proposed to permit the user to specify a presentation. Hence, it is interesting for ScalIntegr to complement and refine my knowledge-presentation related research by investigating (i) a presentation specification language (ontology + commands) and (ii) automatic methods to acquire presentation specifications. This will be useful for the research on "the adaptation of knowledge based on users' profiles" of the KEWI team, or the research on "ergonomic interactive systems" and "models of written and argumentative communication" of the IC3 team. Unlike Fresnel (Pietriga et al., 2006) the presentation specification language will only deal with simple presentation details (such "what kind of information to put in bold characters"), not with graph layout details, but will be integrated in the languages of commands of WebKB-2 and hence the commands will be easy to combine (for example via "pipes" as in Unix) and associate to hyperlinks.
Even for the Semantic Learning Web (Stutt & Motta, 2004) and Educational Semantic Web (Devedzic, 2004) the exploited or acquired user models are very coarse (e.g., grade level and mastery level in certain courses) and most often predefined. The use of a large (virtual) shared KB permits to store what people have seen or been tested on (and when) in a precise and uniform way.
In addition to presenting information in personalised ways, presenting it in an organized way is important too. Indeed, (i) this permit people to see and find what may be of interest to them in a more efficient way, (ii) this may avoid the need to filter information (even manually specified filtering may cause interesting information to be missed, and it is dangerous to let an automatic mechanism guess what is interesting or not for a user), and (iii) this avoids the need to give a list of "related information that might also be of interest". Knowledge normalization guidelines and protocols help achieving or keeping a good organization but additional mechanisms are needed to structure query results. For example, a scalable way to compare objects should be found and developed; an idea to start with is illustrated in Table 2 and presented in (Martin et al., 2005). Long lists of query results should also be structured into conceptual hierarchies.
For this last goal, one approach consists in structuring an initially non-organized long list of statements satisfying a query according to specialization relations that can be calculated or that have been set between the statements. The list thus becomes a list of statement hierarchies the nodes of which a user can navigate, contract or expand. A complementary approach that would require more research and experiments would be to structure the statements of the list according to various topics to which these statements are related. Examples of topics that can be used to group statements about a certain person are: physical characteristics, administrative details, possessions, work related activities, etc. I believe that such groups (and their sub-hierarchies) can be extracted based on either the type of the first relation of each statement, or if this type is too common, the type of the destination concept node of the first relation. It may turn out that be the second relations and their destinations, or other complementary information , will also have to be exploited.
[any Ontolingua, output_language: a KIF, support of: a lexical_search,
part: {a HTML_based_interface, no graph_visualization_interface}];
//the first line of this FCG can be read: any (instance of the KB server) Ontolingua has
// for output_language (an instance of the notation) KIF and is a support of lexical_searches
[any WebKB-2, output_language: {a FCG, a RDF},
part: (a user_interface, part: {an API, a HTML_based_interface,
a CGI-accessible_command_interface,
no graph_visualization_interface}),
support of: (a specialization_structural_retrieval,
kind: {complete_inferencing, consistent_inferencing},
input: (a query, expressiveness: PCEF_logic),
object: (several statement, expressiveness: PCEF_logic)
)]; //"PCEF": positive conjunctive existential formula
|
As previously noted the originality of ScalIntegr's research on (semi-)automatic ontology integration and knowledge extraction from English sentences (KEE) is that both tasks will be conducted jointly and in the context of ScalIntegr's other research tasks (i.e., knowledge integration into a large shared KB, knowledge normalization guidelines and protocols, replication of knowledge between knowledge bases, progressive extension of Formalized English and of the data model of WebKB-2, and support for the collaboratively-built semi-formal states of the art). Conducting these two tasks "jointly" means that they will share similar techniques, that for ontology integration some KEE procedures will be applied on the names, annotations or other informal information associated to conceptual categories, and that the result of the use of these techniques will increase the shared KB and hence will give more background knowledge for these techniques to work. To that end, the integration of ontologies such as FrameNet (an ontology of English verbs), Extended WordNet (WordNet with pre-parsed annotations) and DBpedia (a base of structured information extracted from Wikipedia) will be prime targets. The result will be easily accessible by researchers or applications like the current Multi-Source Ontology of WebKB-2 is today. Finally, as part of the initialization of an ontology for a collaboratively-built semi-formal state of the art in Ontology Engineering, the main techniques of ontology merging and natural language understanding will be represented. Given that (i) all the tasks of ScalIntegr are inter-related, (ii) ScalIntegr starts from my existing work on a shared KB approach and its partial integration of my work for the "semi-independent approach" (WebKB-2 includes WebKB-1), the research presented in this section is an important part of this long term project.
This research fits the "knowledge extraction and integration" and "Models of knowledge and texts" research themes of, respectively, KEWI (I3S laboratory) and IC3 (IRIT laboratory). Furthermore, I3S includes the team "RL" (for Ressources Linguistiques) while Natural Language Understanding (NLU) is a transversal research theme of the IRIT laboratory.
The investigation of semi-automatic ontology integration will start by re-using and extending the techniques and softwares of EXMO for converting knowledge between various languages, aligning ontologies, reconciling discrepancies, performing terminological analysis, etc. Examples of resulting softwares are the Transmorpher 1.0, the API for ontology and the OWL-Lite Alignment tool (Djoufak-Kengue et al., 2007).
Regarding KEE, ScalIntegr will re-use NLU parsers from the INRIA-CNRS UMR7503 (LORIA) but will adapt them to exploit the content of the shared KB and generate knowledge representations that are more precise and normalised than they currently are. For example, the NLU parser should be able to translate "governments should enforce animal rights" into the normalized FE sentence `the enforcement_of_animal_rights_by_governments should happen' the following FE definition for the concept type "enforcement_of_animal_rights_by_governments": `an enforcement with agent some government and with object some animal_rights'. Then, if the user accepts, or if the translation is automatic, the two knowledge representations may be inserted into the KB. Whether actually inserting categories with names such as "enforcement_of_animal_rights_by_governments" is a good idea or not is another research issue; this depends on various factors such as (i) the existence or not of many statements about such a category, (ii) the domain of the KB, and (iii) the possibility or not to later regenerate the definition of the category from its name. Although many sentences will not be fully automatically understandable, normalizing and indexing many of their components (and especially the processes they refer to) will go a long way toward permitting the normalization, interconnection, classification or comparison of statements and other objects. Similarly, since the most important knowledge to extract are basic relations between categories such as subtypeOf, instanceOf, partOf and substanceOf relations, this will be the main goal of ScalIntegr's KEE related investigation. LORIA has many NLU related research projects, also intends to exploit FrameNet, and thus would be a strong partner for this research.
In the KEWI team, bioinformatics related applications are likely to arise. In that context, it is necessary to re-use specialized bioinformatics KEE tools. The Natural Language Processing Group of the Medical Informatics Division of the University Hospital of Geneva have agreed to cooperate with me on extensions of WebKB-2 for bioinformatics related applications and have allowed me to have the sources of GALEN (Trombert-Paviot et al., 2000) which is (one of) the most advanced of medical text parsers.
Overview. My research report and this document have introduced complementary techniques which, when combined, offer a scalable solution to the collaborative building of a "virtual" well-organized KB based on one or several KB servers on the Web or within an intranet or peer-to-peer network (semantic grid). Such KB can be formal (like CYC for a general KB or OpenGALEN for a medical-oriented KB) or semi-formal. It can be used to progressively organize the large amount of unstructured or semi-independently structured information created by communities via their mailing lists, forums, publications or ontologies. With the tools currently used by communities, the organization of the information is very coarse-grained: most often, whole documents or ontologies are indexed (their statements are not organized into one semantic network). Information on what people know, prefer, can provide or have provided is rarely stored. A main goal of KEWI is to address these organization restrictions (and thereby collaboration restrictions) especially for "organizational memories", the "social and collaborative Web" and (e-)learning (Dehors et al., 2006). Among other tools developed by KEWI to achieve that goal is SweetWiki (Buffa et al., 2007) a user-friendly semantic wiki marrying ontologies and folksonomies. This tool is complementary to WebKB-2 which can also be seen as a semantic wiki but only based on a tight semantic network, not folksonomies. Both approaches can and should be integrated. Part of my role in the KEWI team would be to federate, complement and apply its research in the context of projects, especially industrial projects or European projects - and also find or organize such projects. Projects involving corporate memories, bioinformatics applications, (e-)learning applications or tourism-related applications will be prime targets. I have some concrete experience of using WebKB-2 for the last two kinds of applications (Martin, 2005, 2006). Other interesting domains of application that will particularly test the distribution and cooperation approaches described in the previous sections and in my research report will be e-government and e-sciences. In addition to those applications ScalIntegr will permit people to collaboratively build states of the art and complements to Wikipedia in ways that better help researchers, lecturers, students, engineers and decision makers.
Semantically structured complements to Wikipedia for technical or original materials and hotly debated issues. The goals and approach of Wikipedia restrict it to be poorly structured, of a general "encyclopaedic nature" and "consensual". These restrictions are not lessened by the existence of sister-projects (Wiktionary, Wikiquote, Wikibooks, etc.) nor the potential use of Semantic MediaWiki in the future. The approach I propose permits to go beyond these restrictions and, more generally, beyond certain restrictions of the many on-going research works aiming to improve Wikipedia (e.g., they do not fully address the problems related to the existence of a committee or to right of removal by any user, nor do they provide ways to normalize semi-formal semantic networks and keep them well-organized). Hence, ScalIntegr's approach could be used as a complement for Wikipedia. Using fine-grained argumentation structures, a first step is to represent the debates of many Wikipedia pages marked with having a debated content and, at the bottom these pages, will invite people to explore and complement the argumentation structures in WebKB-2. The intellectual challenge and opportunity for recognition may attract a lot of people. This challenge may be viewed as one of the "games masquerading core tasks of weaving the Semantic Web behind on-line, multi-player game scenarios" such as OntoGame (Siorpaes & Hepp, 2007). I have begun the representation of some of these debates (Martin, 2006b).
Semi-formal states of the art. The beginning of Section 2 of my research report lists reasons why storing formal, semi-formal or informal information into personal documents or KBs was a restricting and sub-optimal way of sharing information. Here are some complementary (although related) reasons why the current system of publishing research outputs via informal research articles is sub-optimal for knowledge sharing compared to the collaboratively building of a semi-formal state of the art.
Semi-formal state of the art in Ontology Engineering. No semi-formal state of the art of Knowledge Engineering (KE) currently exists. The need for it is however clear in the KE community. An example of this need is the interest that Michael Denny's "Ontology editor survey" (Denny, 2004) attracted despite the fact that its very coarse level of details made it quite unusable for its main goal: tool comparison. Another example is the Semantic Web Topics Ontology of ISWC 2006. However, this ontology is very poorly structured (it is by no means helping to find a right place or way to insert information) and is only aimed to support the indexation of documents. The initial ontology of KE that I have begun is not an ontology of topics (the classification of topics cannot be normalised, it is task-dependant), it is an ontology of KE tasks, data structures, notations, tools, and their properties (Martin & Eboueya, 2007).
Semi-formal state of the art in Software Engineering. Creating an ontology of Ontology Engineering leads to include many Software Engineering related categories. One of the goals of the KEWI and RAINBOW teams is to represent knowledge about Software Engineering. Hence, the team members of KEWI and I will extend this ontology.
Berners-Lee T., Hendler J. & Lassila O. (2001). The Semantic Web. Scientific American, May 17th, 2001.
Brandom, R. (1998). Action, Norms, and Practical Reasoning. Noûs, Volume 32, Supplement 2, pp. 127-139.
Buckingham-Shum S., Motta E. & Domingue J. (1999). Representing Scholarly Claims in Internet Digital Libraries: A Knowledge Modelling Approach.ECDL 1999 (pp. 423-442), 3rd European Conf. Research and Advanced Technology for Digital Libraries, Paris, France, September 1999.
Buffa M., Gandon F. & Ereteo G. (2007). A wiki on the semantic web. Chapter of a book titled "Emerging technologies for semantic web environments: techniques, methods and applications" and edited by JöRech, Björn Decker and Eric Ras, Fraunhofer of the Institute for Experimental Software Engineering (IESE), Germany, July 2007.
Corby O. & Faron-Zucker C. (2007). Implementation of SPARQL Query Language based on Graph Homomorphism. Proceedings of ICCS 2007, 15th International Conference on Conceptual Structures, Sheffield, UK.
Corby O., Dieng-Kuntz R., Faron-Zucker C. & Gandon F. (2007). Searching the Semantic Web: Approximate Query Processing based on Ontologies. IEEE Intelligent Systems Journal, Vol. 21, No. 1, January/February 2006.
Corcho O. (2005). A layered declarative approach to ontology translation with knowledge preservation. Frontiers in Artificial Intelligence and its Applications. Dissertations in Artificial Intelligence. IOS Press. January 2005.
De Roure D., Jennings N.R. & Shadbolt N.R. (2005). The Semantic Grid: Past, Present and Future. Proceedings of the IEEE, vol. 93, no.3, pp. 669-681, March 2005.
Dehors S., Faron-Zucker C. & Dieng-Kuntz R. (2006). Reusing Learning Resources based on Semantic Web Technologies. Proceedings of ICALT 2006, 6th IEEE International Conference on Advanced Learning Technologies, Kerkrade, The Netherlands, July 2006.
Denny M. (2004). Ontology Tools Survey, Revisited. Web document created on July 14th 2004. http://www.xml.com/pub/a/2004/07/14/onto.html
Devedzic, V. (2004). Education and the Semantic Web. International Journal of Artificial Intelligence in Education, 14, pp. 39-65.
Dieng-Kuntz R. & Corby O. (2005). Conceptual Graphs for Semantic Web Applications. "Conceptual Structures: Common Semantics for Sharing Knowledge", LNCS 3596/2005, pp. 19-50.
Djoufak-Kengue J.-F., Euzenat J. & Valtchev P. (2007). OLA in the OAEI 2007 evaluation contest. Proceedings of ISWC 2007 workshop on ontology matching, pp. 117-184, Busan, Korea.
Ellis G. (1995). Compiling Conceptual Graphs. IEEE Transactions on Knowledge and Data Engineering, February 1995, Vol. 7, No. 1, pp. 68-81.
Euzenat J. & Shvaiko P. (2007). Ontology Matching. Springer-Verlag, Berlin Heidelberg (DE), 2007, 341 pages.
Friedland N.S., Allen P., Mathews G., Whitbrock M., Baxter D., Curtis J., Shepard B., Miraglia P., Angele J., Staab S., Moench E., Opperman H., Wenke D., Israel D., Chaudhri V., Porter B., Barker K., Fan J., Chaw S.Y., Yeh P., Tecuci D. & Clark P. (2004). Project Halo: Towards a Digital Aristotle. AI Magazine, 25(4), 2004, pp. 29-48.
Hillis W.D. (2004). "Aristotle" (The Knowledge Web). Edge Foundation, Inc., No 138, May 6, 2004.
Kalfoglou Y., Alani H., Schorlemmer M. & Walton C. (2004). On the Emergent Semantic Web and Overlooked Issues. Proceedings of ISWC'04, 3rd International Semantic Web Conference, LNCS 3298, Hiroshima, Japan, pp. 576-, November 2004.
Martin Ph. (1993). A KADS refinement for Explanatory Knowledge Extraction and Modelling. Proceedings of AI 1993, 6th Australian Joint Conference on Artificial Intelligence (edited by "World Scientific, Singapore"), Melbourne, Australia, November 16-19, 1993.
Martin Ph. (1996). Exploitation de graphes conceptuels et de documents structurés et hypertextes pour l'acquisition de connaissances et la recherche d'informations. Ph.D. thesis (378 pages), University of Nice - Sophia Antipolis, France, October 14, 1996.
Martin Ph. (2000). Conventions and Notations for Knowledge Representation and Retrieval. Proceedings of ICCS 2000, 8th International Conference on Conceptual Structures (Springer, LNAI 1867, pp. 41-54; electronically published on 1/1/2007), Darmstadt, Germany, August 14-18, 2000.
Martin Ph. & Eklund P. (2001). Large-scale cooperatively-built heterogeneous KBs. Proceedings of ICCS 2001, 9th International Conference on Conceptual Structures (Springer, LNAI 2120, pp. 231-244), Stanford University, California, USA, July 30 to August 3, 2001.
Martin Ph. (2002). Knowledge representation in CGLF, CGIF, KIF, Frame-CG and Formalized-English. Proceedings of ICCS 2002, 10th International Conference on Conceptual Structures (Springer, LNAI 2393, pp. 77-91), Borovets, Bulgaria, July 15-19, 2002.
Martin Ph. (2003). Correction and Extension of WordNet 1.7. Proceedings of ICCS 2003, 11th International Conference on Conceptual Structures (Springer, LNAI 2746, pp. 160-173), Dresden, Germany, July 21-25, 2003.
Martin Ph., Blumenstein M. & Deer P. (2005). Toward cooperatively-built knowledge repositories. Proceedings of ICCS 2005, 13th International Conference on Conceptual Structures (Springer, LNAI 3596, pp. 411-424), Kassel, Germany, July 18-22, 2005.
Martin Ph. (2005). Services on the Sunshine Coast. http://www.webkb.org/SC/
Martin Ph. (2006). Documents related to my Griffith E-Learning Fellowship for Semester 2, 2006. http://www.webkb.org/doc/papers/GEL06/
Martin Ph. (2006b). Structured discussions & Semantic classification of some resources. http://www.webkb.org/kb/it/
Martin Ph., Eboueya M., Blumenstein M. & Deer P. (2006). A Network of Semantically Structured Wikipedia to Bind Information. Proceedings of E-learn 2006 (pp. 1684-1702), AACE Conference on E-learning in Corporate, Government, Healthcare, & Higher Education, Honolulu, Hawaii, October 13-17, 2006.
Martin Ph. & Eboueya M. (2007). Sharing and Comparing Information about Knowledge Engineering. WSEAS Transactions on Information Science and Applications, Issue 5, Volume 4 (pp. 1089-1096; ISSN: 1790-0832), May 2007.
Martin Ph. & Eboueya M. (2008). For the ultimate accessibility and re-usability. Book chapter to be published early 2008 in the Handbook of Research on Learning Design and Learning Objects: Issues, Applications and Technologies. -->
Messmer B.T. & Bunke H. (2000). Efficient Subgraph Isomorphism Detection: A Decomposition Approach. IEEE Transactions on Knowledge and Data Engineering, March 2000, Volume 12, Issue 2, pp. 307 - 323.
Noy N.F., Chugh A., Liu W. & Musen M. A. (2006). A framework for ontology evolution in collaborative environments. Proceedings of ISWC 2006, 5th International Semantic Web Conference, Athens, GA, 2006. ICCS 2003, 11th International Conference on Conceptual Structures (Springer Verlag, LNAI 2746, pp. 362-377), Dresden, Germany, July 21-25, 2003. -->
Patel-Schneider P.F. (2005). A Revised Architecture for Semantic Web Reasoning. Proceedings of Third Workshop on Principles and Practice of Semantic Web Reasoning (PPSWR'05), Dagstuhl, Germany (11th - 16th September 2005), LNCS 3703, pp. 32-36, September 2005.
Pietriga E., Bizer C., Karger D., Lee R. (2006). Fresnel: A Browser-Independent Presentation Vocabulary for RDF. Proceedings of ISWC 2006, 5th International Semantic Web Conference, LNCS 4273, pages 158-171, November 2006, Athens, GA, USA.
Rousset M-C. (2004). Small Can Be Beautiful in the Semantic Web. Proceedings of International Semantic Web Conference (ISWC 2004), pp. 6-16.
Shadbolt N., Berners-Lee T. & Hall W. (2006). The Semantic Web Revisited. IEEE Intelligent Systems, 21(3) pp. 96-101, May/June 2006.
Siorpaes K. & Hepp M. (2007). OntoGame: Towards Overcoming the Incentive Bottleneck in Ontology Building. Proceedings of the 3rd International IFIP Workshop On Semantic Web & Web Semantics (SWWS 2007), LNCS 4806, pp. 1222-1232, November 29-30, 2007.
Stutt A. & Motta E. (2004). Semantic Learning Webs. Journal of Interactive Media in Education, Special Issue on the Educational Semantic Web, 10.
Trombert-Paviot B., Rodrigues J.M., Rogers J.E., Baud R., van der Haring E., Rassinoux A.M., Abrial V., Clavel L. & Idir H. (2000). GALEN: a third generation terminology tool to support a multipurpose national coding system for surgical procedures. International Journal of Medical Informatics, Vol. 58-59 (pp. 71-85), September 1st 2000.
Vrandecic D., Pinto H.S., Sure Y. & Tempich C. (2005). The DILIGENT Knowledge Processes. Journal of Knowledge Management, 9(5), pp. 85-96, October 2005.
Weitzner D. (2007). Reciprocal Privacy (ReP) for the Social Web. Web document created on December 12th 2007. http://dig.csail.mit.edu/2007/12/rep.html
Yao H. & Etzkorna L. (2006). Automated conversion between different knowledge representation formats. Knowledge-Based Systems, Volume 19, Issue 6, October 2006, pp. 404-412.