CAB: A tool for interoperation among cognitive architectures

Abstract

The rapid technology development for modeling and simulating human behavior and cognition as Intelligent Virtual Agents (IVAs) has resulted in broad incompatibilities among underlying architectures and specific models. At the same time, the growing interest in practical application of IVAs in defense/aerospace, healthcare, and training systems is bringing demands of easier and cheaper IVA development, and increased re-usability of agent features and components. The Cognitive Architecture Bridge (CAB) is being developed as a new 'middleware' approach to providing multiple levels of interoperability and composibility between and among IVA models and model-components. CAB is designed to allow capabilities from different IVA models and modeling architectures to be integrated into a single 'virtual' IVA model, through four general classes of mechanisms and a common run-time infrastructure. A procedure call mechanism allows one IVA model to call a procedure from another model, e.g., to perform a specific cognitive task, or to perform an action on the external world. This is a synchronous form of interoperation, where the calling model thread waits for the completion of the function before continuing its own thread of execution. A message passing mechanism allows information to be sent explicitly from one model to another, in the form of a message which may contain any type of data, not only textual information. Unlike procedure calling, this is an asynchronous interoperation mechanism, in which the sending model thread continues its execution without waiting for feedback from the model receiving the message. A distributed memory mechanism enables parts of the memory content of one model to become visible to other models or components of other models. Unlike the procedure call and message passing interaction modes, the distributed memory mode does not require any explicit invocation, but instead allows a model to react to changes in the memory state of another model just as it would react to changes that occur in its own memory contents. Finally, a resource sharing mechanism allows an architectural construct or component to be used by many other components or models, but only by one entity at a time. For example, an eye resource could be used by several partial IVA models to provide vision, but could only be controlled by one of them at a time. A preemption capability is included to reassign the resource to a model having a suddenly higher- priority need for the resource. Existing IVA engines and architectures can be made CAB-compliant through a one-time adaptation to support the CAB application program interfaces. © Springer-Verlag Berlin Heidelberg 2006.