Learning and Planning of Situated Resource Bounded Agents
Year: 2006
Type: conference-paper
Abstract: This paper presents an investigation of rational agents that have limited computational resources and that can interact with their environments. We analyse how such agents can combine deductive reasoning using domain knowledge and inductive learning from past experiences, while remaining time-aware in a manner appropriate for beings situated in a dynamic universe. In particular, we consider how they can create and reason about partial plans, choose and execute the best ones of them — in such way as to acquire the most knowledge. We also discuss what are the different types of interactions with the world and how they can influence agent’s ability to consciously direct its own learning process.
Author: Sławomir Nowaczyk
Institution: Lund University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1745391901