인공지능/Artificial Intelligence (COSE361)19 [인공지능] Chapter 8. Adversarial Search (2) Uncertain outcomes controlled by chance, not an adversary! Expectimax Search 어떠한 행동의 결과가 어떻게 될지 알 수 없는 이유는? Explicit randomness Unpredictable opponents Actions can fail Values should now reflect average-case (expectimax) outcomes, not worst-case (minimax) outcomes Expectimax search Compute the average score under optimal play Minimax search와 같은 max node Chance node are like min nodes but the out.. 2023. 10. 25. [인공지능] Chapter 7. Adversarial search (1) Games There are many different kinds of games Deterministic or stochastic One, two, or more players? Zero sum? Perfect information (can you see the state)? We want algorithms for calculating a strategy (policy) which recommends a move from each state Deterministic Games States Players Action Transition function Terminal test Terminal Utilities Zero-sum games Zero-sum games Agents는 서로 반대되는 utilit.. 2023. 10. 25. [인공지능] Chapter 6. Constraint Satisfaction problems What is search for? Assumptions about the world a single agent deterministic actions fully observed state discrete state space Planning: Sequences of actions Goal에 이르는 Path가 중요 Paths는 다양한 비용과 깊이(depths)를 가짐 Heuristics give problem-specific guidance Identification: assignments to variables Goal 그 자체가 중요 (Path가 아니라) 모든 Paths는 같은 depth를 가짐 CSPs are a specialized class of identification problems Con.. 2023. 10. 24. [인공지능] Chapter 5. Propositional logic Propositional Logic 명제 논리 Basic concepts of knowledge, logic, reasoning Propositional logic: syntax(구문) and semantics(의미론), Pacworld example Inference by model checking and theorem proving Agents that know things Agents는 지각, 학습, 언어를 통해 지식을 습득 Knowledge of the effects of actions trainsition model Knowledge of how the world affects sensors sensor model Knowledge of the current state of the world C.. 2023. 10. 24. [인공지능] Chapter 4. Local Search Local Search Algorithms 많은 최적화 문제에서, paths는 irrelevant하다. The goal state is the solution Then state space = set of ‘complete’ configurations; find configuration satisfying constraints (n-queens problem) or find optimal configuration (travelling salesperson problem) In such cases, we can use iterative improvement algorithms keep a single ‘current’ state, try to improve it constant space, suitable.. 2023. 10. 24. [인공지능] Chapter 3. Informed Search Heuristic 특정 state가 얼마나 goal state에 가까운지 평가하는 함수 특정 Search problem을 위해 고안됨 Greedy Search goal state에 가장 가까워 보이는 노드로 확장 Strategy expand a node that you think is closest to a goal state Heuristic 활용: 각각의 state에서 가장 가까운 goal까지의 거리 계산 A common case Best-first takes you straight to the goal Worst-case kike a badly-guided DFS A* Search UCS + Greedy Uniform-cost orders by path cost, or backward cost g(.. 2023. 10. 24. [인공지능] Chapter 2. Search Agent Reflex agents Current percept 기반 행동 행동의 결과를 고려하지 않음 Consider how the world IS Planning agents 행동의 결과를 고려 Consider how the world WOULD BE Search Problems Consist of A state space 모든 가능한 상태들의 집합 A successor function 현재 상태를 기반으로 다음 상태 및 비용을 연산하는 함수 A start state 초기상태 A goal test 주어진 상태가 목적 상태인지 확인하는 함수 Solution start state를 goal state로 바꾸는 sequence of actions State space Graphs A mathmatical .. 2023. 10. 24. 이전 1 2 다음
