The goal of the GBSML project is to support the rapid development and deployment of web-based interactive learn by doing education and training environments, using distributed Semantic Web content to provide centralized reusable subject matter content. This project will develop an ontological vocabulary of training scenarios, using the general Goal-Based Scenario (GBS) architectures developed and applied over the past decade. More specifically, we are developing (1) the Goal-Based Scenario Markup Language (GBSML), a DAML-compatible markup language for describing domain content in a form suitable for supporting education and training over the Semantic Web, (2) several prototype web-based learn-by-doing goal-based scenario applications using example GBSML content, (3) web services and tools to generate and distribute GBSML content from HTML, XML, and GBSML sources.
1.1 Procedural scenarios: We are developing GBSML markup for training scenarios involving multi-step procedures. This requires an ontology involving some planning concepts, such as actions, effects, preconditions, and goals, but extended to provide markup about mistakes and their effects (steps missed, steps done in the wrong order, actions applied to the wrong objects). Tasks and deliverables:
1.1.1 GBSML 1.0: A markup language for describing procedures, including correct and incorrect sequences of actions, action prerequisites and consequences, warnings, and so on.
1.1.2 GBSML-based Procedure Trainer: A run-time engine to deliver web-based interactive training for procedures described in GBSML.
1.1.3 Sample Trainer: A small but real sample procedural GBS trainer.
2.1 Analytical Report Scenarios: We will extend GBSML to support for scenarios involving critical reasoning and argumentation skills. This requires an ontology of arguments, i.e., claims and evidence, extended to provide markup for weaknesses in arguments (inconsistent or incomplete evidence, alternatives not disproven, and so on). Tasks and deliverables:
2.1.1 GBSML 2.0: GBSML extended to describe evidential relationships between observations, hypothesized situations, and courses of action, capable of supporting automated critiquing of argument-based reports and plans.
2.1.2 GBSML-based Argument Critiquer: A run-time engine capable of providing web-based critiquing of argument-based reports, using GBSML 2.0 content .
2.1.3 Sample Critiquer: A small but real sample argument critiquer.
Key metrics for measuring progress in the GBSML project are:
· Publication and review of the initial versions of the GBSML ontologies.
· Application of the ontologies to the markup of short documents describing simple but realistic examples of (1) multi-step procedures (2) analytical claim and evidence relationships.
· Implementation of web-based proof-of-concept training systems using GBSML marked-up content.
The focus of the GBSML project is to develop ontologies and web services to support anywhere anytime small-scale training modules. The DAML Experiment, on the other hand, is intended to demonstrate support for the operational military commander. The Demonstration is focused on the Knowledge aspects of the JFCOM Rapid Decisive Operations and specifically Operational Net Assessment (ONA).
As such, there is no direct connection between the experiments relevant to GBSML systems and those in the current DAML Demonstration.