Graphical ontology modeling language for learning environmentsIn the last fifteen years, our main goal has been to synthesize and combine various forms of graphical representations that are useful for educational modeling and knowledge management, using an integrated graphical formalism. We have shown that very different kinds of representation, conceptual maps, flowcharts, decision trees and others, can all be modeled more precisely, using the MOT graphic language based on typed objects (concept, procedures, principles, facts) as well as few typed links. With this set of primitive graphic symbols, it has been possible to build very different graphic models, from simple taxonomies to ontologies, more or less complex learning designs, delivery process, decision systems, and methods.Recent developments have led to two specialisations of the graphic language. The first one is a powerful, yet simple graphic language to build ontologies for a knowledge domain. The second one enables to model learning designs and scenarios in a standardized and computable way. The association between both kinds of models specifies the central part of a learning environment at the design phase, and enables its delivery to learners and educators. In the final section, I assert that knowledge representation for education should be graphic, user-friendly, general, scalable, declarative, standardized and computable.Gilbert PaquetteGP_007c2d10-dc9e-11ec-8f85-34cb2c83ea00Author | Télé-universitéGraphical representations for educational modeling and knowledge management_007c2ebe-dc9e-11ec-8f85-34cb2c83ea00To support the design and delivery of learning and knowledge-intensive environments that are technology-enhanced_007c3094-dc9e-11ec-8f85-34cb2c83ea00Knowledgeknowledge representation should be graphic, user-friendly, general, declarative, standardized and computableTechnologyRelationshipsGraphic ~ The benefits of graphical cognitive modelling have been eloquently summarized by many authors, Ausubel (1968), Dansereau (1978) and Jonassen (1993) to name a few. Graphical models illustrate relationships among components of complex phenomena.InteractionThey uncover the complexity of actors’ interactions.CommunicationThey facilitate the communication about the reality studied.ComprehensionThey favour the global comprehension of studied phenomena. They help grasp the structure of related ideas by minimizing the use of ambiguous natural language texts. User-FriendlinessIn our view, a representation system intended to support pedagogy must be easy to use without technical or scientific mastery after a relatively short period of initiation. Dansereau and Holley, (1982) have studied experimentally the use of different sets of graphic symbols by learners. Their results show that typed links are preferred by the majority of learner, as long as there are not two few nor two many links. In other words, the components of a graphic system must be easy to interpret. The meaning of links between knowledge elements must be sufficiently distinct from one another, yet they must capture natural ways of thinking. The number and variety of MOT applications in the last fifteen years qualify on that account.GeneralityGenerality means that the representation language should have the capacity to represent, with a relatively small number of objects and link categories, all knowledge in very different subject domain, at various levels of granularity and formality. ScalabilityThe graphic language should be scalable from informal graphs, up to semi-formal and totally unambiguous formal models.DeclarationDeclarative ~ Graphic language can be procedural or declarative. There have been many discussions in the past on that issue. Procedural graphic languages have been built in the past; essentially extending flowcharts to promote graphical programming that would produce code directly. Our proposal is to use, as much as possible, a declarative graphic language, for a number of reasons. Firstly, it is easier for a person to declare the components of his/her knowledge than to describe also the way it should be processed. Procedural languages mix up the knowledge with the processing on the knowledge, blurring the interpretation of the knowledge representation. In expert systems for example, the execution instructions are not wired-in the program, but externalized and made visible in a knowledge base on which a general inference engine proceeds.ReusabilitySecondly, the same model, if declarative, can be used more easily for different applications, not necessarily the one for which the processing has been planned in a procedural program. For example, rules for operating or for diagnosing a component-based system can be applied to different conceptual models describing a car, a software system or a learning environment. This can be done by querying the model using an inference engine, in a Prolog-like manner. Ontology-driven architecture, a promising approach, is based on that very idea.AnalysisThirdly, the processing knowledge itself can be given declaratively, so that higher order meta-knowledge, can be also singled-out. This idea is similar to structural analysis (Scandura 1973) and it is exactly the way we should see the relation between generic skills and domain knowledge in a competency.SkillA generic skill, like apply, synthesize or evaluate is meta-knowledge applied to domain knowledge, stating for example that a certain concept can be applied, synthesized or evaluated by somebody. Many generic skills have been described by MOT process-model, thus providing skeletons of activities where the structure of the skill can guide its application to specific domain knowledge (Paquette 1999).SynthesisEvaluationInferenceThis is similar to the way the KADS methodology propose to operate when task and inference models are applied to domain knowledge (Breuker and Van de Velde, 1994).StandardizationStandardization is an important property to enlarge knowledge use and communication between users, persons or software agents. At the informal level, each model constructed by a person must be interpretable by another person. At the formal level, the communication capabilities extend to software agents. This is why we have standardized the MOT representation system as a tool within our research organization, facilitating the communication between participants in sometimes very different projects. The move towards graphic versions of standards like IMS-LD for learning designs and OWL-DL for ontologies adds wider communication capabilities between researchers and educators while at the same time adding formal non-ambiguous interpretation for machine processing.ComputationComputability is a step ahead from standardization. It means not only that the graphic model can receive a non-ambiguous formal representation that can be processed by computer agents (this is the case of MOT+LD models), but that this formal representation is complete, that is all conclusions are guaranteed to be computable, and decidable, that is all computations will finish in finite time. These considerations have motivated the construction of the MOT+OWL graphic language as an equivalent to the OWL-DL language-based on description logic while much simpler and user friendly than other editors we are aware of. Base on this standard, the use of models built with MOT+OWL will benefit from the growing number of tools and applications that are and will be developed by the international community.DesignSupport the design of learning environments_007c3224-dc9e-11ec-8f85-34cb2c83ea001Learning Environment DesignersLearners[C]ompetency annotations of the activities, the resources and the actors in a Learning Design can be used ... to support the design and delivery of learning environments. At design time, they can help designers (or learners acting as their own designers) to:Learning UnitsPrepare a sequence of learning units that should increase progressively the mastery level of learners_007c333c-dc9e-11ec-8f85-34cb2c83ea001.1ResourcesIdentify learning resource (documents, tools, activities, persons) to be included in a learning design that possess the right knowledge and the right mastery level to help learner progress_007c3454-dc9e-11ec-8f85-34cb2c83ea001.2CriteriaProvide criteria to form teams with learners having homogenous or heterogeneous mastery levels_007c3576-dc9e-11ec-8f85-34cb2c83ea001.3LearnersLearning TeamsPaths & PlaysPlan different paths or plays for learners with weak or strong mastery patterns_007c38dc-dc9e-11ec-8f85-34cb2c83ea001.4DeliverySupport the delivery of learning environments_007c3a08-dc9e-11ec-8f85-34cb2c83ea002At delivery time, competency notations can help learners and trainers to:EvaluationEvaluate the progress of learners’ competencies for important knowledge elements_007c3b48-dc9e-11ec-8f85-34cb2c83ea002.1RiskDetect learners at risk by comparing their evolution pattern to the group average_007c3c88-dc9e-11ec-8f85-34cb2c83ea002.2FlawsAlert learners, trainers and designers on possible flaws in the learning environment_007c3daa-dc9e-11ec-8f85-34cb2c83ea002.3LearnersTrainersDesignersResourcesFind appropriate resources or units of learning in a learning object repository where resources have been referenced with semantic annotations_007c3ee0-dc9e-11ec-8f85-34cb2c83ea002.4User ModelsBuild/maintain user models_007c3fee-dc9e-11ec-8f85-34cb2c83ea002.5Learning System UsersInterventionsGuide trainer interventions_007c411a-dc9e-11ec-8f85-34cb2c83ea002.5.1TrainersAdvice & TutoringTrigger intelligent advisors and tutoring systems_007c423c-dc9e-11ec-8f85-34cb2c83ea002.5.2E-PortfoliosAdd information e-portfolio systems_007c4354-dc9e-11ec-8f85-34cb2c83ea002.5.32022-05-25https://www.researchgate.net/publication/228639039_Graphical_ontology_modeling_language_for_learning_environmentsOwenAmburOwen.Ambur@verizon.net