Output details
11 - Computer Science and Informatics
University of Edinburgh
Temporal Vagueness, Coordination and Communication
<22> Originality: This paper develops a novel synthesis of agent-based approaches to ontology alignment, reinforcement learning, the semantics and pragmatics of approximate temporal terms, Gardenfors-style conceptual spaces, and a utility-based model of vagueness.
Significance: Provides a multi-agent simulation of task-based communication language game in the sense of Gardenfors, where the representations of individual agents affect communication about shared activities and are modified as a result. The paper argues that a notion of approximate semantic alignment may be sufficient for communication in a task-oriented scenario.
Rigour: The approach is tested by detailed simulations of agent-based interaction under a number of conditions.