Is Everything Going According to Plan? Expectations in Goal Reasoning Agents

In part motivated by topics such as agency safety, there is an increasing interest in goal reasoning, a form of agency where the agents formulate their own goals. One of the crucial aspects of goal reasoning agents is their ability to detect if the execution of their courses of actions meet their own expectations. We present a taxonomy of different forms of expectations as used by goal reasoning agents when monitoring their own execution. We summarize and contrast the current understanding of how to define and check expectations based on different knowledge sources used. We also identify gaps in our understanding of expectations.

Rapid prototyping of cognitive agent simulations using C-BML transformations

Simulating battle management is an essential technique used in planning and mission rehearsal as well as training. Simulation development costs tend to be high due to the complexity of cognitive system architectures in such applications. Due to this complexity, it takes significant effort for a simulation developer to comprehend the problem domain enough to capture accurately in a simulation code. Domain-specific languages (DSL) play an important role in narrowing the communication gap between the domain user and the developer and hence facilitate rapid development. In command and control (C2) applications, the coalition battle management language (C-BML) serves as a DSL for exchanging battle information among C2 systems, simulations, and autonomous elements. In this article, we use a rapid prototyping framework for cognitive agents and demonstrate deployment of agent systems by adopting the model driven engineering approach. To this end, we extend the use of C-BML and automatically transform it in a cognitive agent model, which is then used for adaptive decision making at runtime. As a result, during a simulation run, it is possible to initialize and modify an agent’s goal reasoning model. The cognitive agent models are based on the deliberative coherence theory, which provides a goal reasoning system in terms of coherence-driven agents.

Goal Reasoning: Foundations, Emerging Applications, and Prospects

Goal reasoning (GR) has a bright future as a foundation for the research and development of intelligent agents. GR is the study of agents that can deliberate on and self-select their goals/objectives, which is a desirable capability for some applications of deliberative autonomy. While studied in diverse AI sub-communities for multiple applications, our group has focused on how GR can play a key role for controlling autonomous systems. Thus, its importance is rapidly growing and it merits increased attention, particularly from the perspective of research on AI safety. In this article, I introduce GR, briefly relate it to other AI topics, summarize some of our group’s work on GR foundations and emerging applications, and describe some current and future research directions.