Toward Meta-level Control of Autonomous Agents
نویسندگان
چکیده
Metareasoning is an important capability for autonomous systems, particularly for those being deployed on long duration missions. An agent with increased self-observation and the ability to control itself in response to changing environments will be more capable in achieving its goals. This is essential for long-duration missions where system designers will not be able to, theoretically or practically, predict all possible problems that the agent may encounter. In this paper we describe preliminary work that integrates the metacognitive architecture MIDCA with an autonomous TREX agent, creating a more self-observable and adaptive agent.
منابع مشابه
Attitudes for Agents in Dynamic Worlds
In this paper, we propose that in multiagent dynamic worlds, agents which are autonomous, need to be guided by attitudes for effective problem solving. Often agents, existing in dynamic and hostile worlds, are required to exhibit the specified problem solving behaviors for prolonged periods of time either continually or intermittently. In order to be able to perform these types of behaviors, au...
متن کاملThe Use of Meta-Level Control for Coordination in a Distributed Problem Solving Network
Distributed problem-solving networks provide an interesting application area for meta-level control through the use of organizational structuring. We describe a decentralized approach to network coordination that relies on each node making sophisticated local decisions that balance its own perceptions of appropriate problem-solving activity with activities deemed important by other nodes. Each ...
متن کاملA Framework for Autonomous Agents Based on the Concept of Anticipatory Systems
This paper presents a new framework for autonomous agents that is based on the concept of anticipatory systems. It is a hybrid approach that synthesizes low-level reactive behavior and high-level symbolic reasoning. According to this framework, an agent, i.e. an anticipatory agent, consists of three main entities: a reactive system, a world model, and a meta-level component. The world model sho...
متن کاملShout with the largest Mob : toward a model for primitive communication in mobile automata
We consider the problem of simulating communication between independent, autonomous agents, or machines, using only local rules with no global control over the agents' behavior. First, we construct an algorithm by which the machines will avoid collisions with each other and with boundaries or obstacles. Noting that collision avoidance alone will not result in higher-level behavior, and with the...
متن کاملMeta-Learning with Backpropagation
This paper introduces gradient descent methods applied to meta-leaming (leaming how to leam) in Neural Networks. Meta-leaning has been of interest in the machine leaming field for decades because of its appealing applications to intelligent agents, non-stationary time series, autonomous robots, and improved leaming algorithms. Many previous neural network-based approaches toward meta-leaming ha...
متن کامل