AI-ordlista
Den kompletta ordlistan över AI
Game Theory in Multi-Agent Learning
Application of mathematical game theory models to analyze and design learning strategies in environments where agents have interdependent interests.
Nash Equilibrium in Learning
Concept where no agent can improve its reward by unilaterally changing its strategy, used as a convergence criterion for multi-agent learning algorithms.
Competitive Learning
Learning process where agents develop strategies in direct competition for limited resources or objectives in a shared environment.
Cooperative Learning
Approach where agents collaborate to achieve a common goal, often sharing information or coordinating their actions to optimize a collective reward.
Learning by Emulation
Technique where agents learn by imitating successful behaviors of other agents observed in the environment, thus accelerating the skill acquisition process.
Stability in Multi-Agent Learning
Property guaranteeing that learning policies converge to a predictable equilibrium state despite dynamic interactions between agents.
Exploration vs Exploitation in Multi-Agents
Complexified dilemma where each agent must balance discovering new strategies with using existing knowledge, while anticipating the choices of other agents.
Dynamic Task Allocation
Learning process where agents negotiate and adapt to efficiently distribute changing tasks in a multi-agent environment.
Policy Convergence
Phenomenon where agents' strategies stabilize toward a set of coherent policies after a period of learning and mutual adaptation.
Multi-Agent Trial and Error Learning
Methodology where agents explore the space of possible actions and adjust their behaviors based on observed successes and failures in a multi-agent context.