Centralized Knowledge Representation
In most large, distributed systems that use state-of-the-art components, there invariably exist mismatches in how these components internally represent concepts. Without due care, distributed resources' representations can get out of synch, contain localized errors, or become manageable only by a small group of experts for each module. While there was no reason during initial development of the Virtual Humans in MRE and SASO for the various modules to be uniform or employ exactly the same formats of knowledge representation, as the Virtual Human matures, there is an increasing need for standardization.
The aims of this part of the project are to create a framework for the system that can embed versions of modules and code that are standardized, well-integrated, easily extensible, and reusable in other projects. Centrally, this new single framework holds the core knowledge representation structures of all the principal modules in one shared space, and supports tools and interfaces that allow system builders to view, edit, debug, and extend the knowledge representations uniformly for all modules at once. The centralized knowledge repository includes an ontology (for concept and instance definitions), a task/plan/script library, natural language processing information in lexicons, a framebank of semantic frame representations and their associated surface realizations, and other information. The framework also supports the automated consistency checking of changes and the propagation of new knowledge to all relevant modules in the appropriate formats.
This project differs from others in the following ways:
- It includes both understanding and generation (most projects focus on one direction only)
- It includes prosodic information for parsing long sentences (no other project we know of does this)
- It explicitly combines statistical and rule-based components (most projects take one or the other approach exclusively)
- It requires less training data than most projects of this kind, because of the rule-based methods employed
- It is more ambitious than most others in the field in that the Virtual Humans contain more modes of interaction (gesture, speech, language, etc.); correspondingly, the central knowledge representation framework is more complex than most
Tags: human, knowledge, language, natural, representation, symbols, understanding, virtual
Goals
- Modeling knowledge such that it can be shared by all modules
- Modeling knowledge such that it can be reused for new scenarios
- Providing tools to allow developers to both create new and tweak existing scenarios, supporting the preferred activities of power users specialized in the system as well as domain experts with less system expertise