Thuật ngữ AI
Từ điển đầy đủ về Trí tuệ nhân tạo
Distributed reasoning
Process by which multiple agents collaborate to reach a logical conclusion, each contributing with their own inferences and local knowledge without centralized supervision.
Distributed constraint satisfaction
Approach where agents collaborate to find a solution satisfying a set of interdependent constraints, each agent managing its own variables and local constraints.
Distributed reinforcement learning
Paradigm where multiple agents simultaneously learn optimal policies by sharing their experiences and rewards, thus accelerating convergence toward collective solutions.
Hierarchical problem decomposition
Technique consisting of dividing a complex problem into simpler sub-problems organized in a hierarchical structure, each level being managed by different agents or groups of agents.
Distributed information fusion
Process of aggregating and synthesizing data collected by multiple agents to produce global knowledge that is more complete and accurate than that of an isolated agent.
Distributed planning
Method where agents collaborate to develop a common action plan by coordinating their local activities and resolving objective or resource conflicts.
Collaborative problem solving
Systemic paradigm where agents actively share their knowledge, strategies, and partial results to progressively build a solution to the global problem.
Emergence of collective solutions
Phenomenon by which complex and non-trivial solutions spontaneously arise from local interactions between agents, without explicit centralized design of the global solution.
Contract-based coordination
Organizational mechanism where agents establish formal agreements to define their mutual responsibilities in the distributed resolution of a complex problem.
Distributed resilience
Capacity of a multi-agent system to maintain its problem-solving functionality despite the failure of certain agents, through redundancy and dynamic reorganization.
Adaptive agent specialization
Process by which agents spontaneously develop complementary expertise to collectively optimize problem resolution, based on the system's needs.
Goal-oriented communication
Information exchange protocol between agents optimized for problem resolution, where messages are filtered and prioritized based on their relevance to the common objective.