Intelligent Agents

Agents and Environments

Agent 可以是任何一種東西，用 censor 來感測環境，並且透過 actuators 來動作

• An agent’s behavior is described by the agent function that maps any given percept sequence to an action.

  • e.g., vacuum-cleaner

    • if the current square is dirty, then suck

    • otherwise, move to the other square

Rational agent

• Rational agent 指的是可以做出正確事情的 agent

• 正確事情 : Agent 所做的事情會讓 enviroment 歷經一系列的 desirable states

• desirability : performance measure 來量測

• Definition : Rational agent 會在 percept 的每個步驟都需要 maximize performance measure，從而 action

• Example : vacuum-cleaner agent

  • Performance measurement : the amount of dirt being cleaned up

  • Performance measure: a clean floor

Omniscience, Learning, and Autonomy

Rationality maximizes expected performance, while perfection maximizes actual performance.

• 針對 Agent 的 rationality 不需要達到 omniscience

• 但 Agent 需要能夠從 perceive 的 data 中學習 (learn)

• 另外 Agent 最好能夠 Autonomous

  • learn and compensate for incorrect

Task Environments

我們透過以下幾點，來定義一個 Task environment :

• Performance measure

• The environment

• Agent's actuators

• Agent's sensors

• 合稱為 PEAS (Performance, Environment, Actuators, Sensors)

例如 :

Agent	Performance	Environment	Actuators	Sensors
Taxi driver	Safety, fast	Traffics, Customers	Brake, Accelerator	GPS, Engines

Properties of Task Environments

Fully observable vs. partially observable

• Agent 的每個 action 是否會跟 sensors 偵測所有的東西有關

Single agent vs. multiagent

• 如果有多個 Agents，那他們之間是否會互相影響對方

• Chess 就是 multiagent

Deterministic vs. stochastic

• 若 environment 的下一個 state 會受前一個 state 影響，那就是 deterministic

• stochastic 則相反，通常代表 outcomes 是 uncertainty 的

• Taxi driving 就是 stochastic

Episodic vs. sequential

• 情節之間是否是連貫的

• Chess & Taxi driving 都是 sequential

Static vs. dynamic (taxi)

• Environment 會因為 Agent 的思考而改變，則說 environment 是 dynamic 的

• Taxi driving 是 dynamic 的

Discrete vs. continuous

• Chess 的 environment 有 finite 的 discrete states & actions

• Taxi driving 則是有 continuous state

Known vs. unknown

• The agent's state of knowledge about the “laws of physics” of the environment.

The Structure of Agents

• Agent = Architecture + Program

• Program : 設計來 implement agent function (用於 map percepts to action)

• Architecture : 指的是 Devices (sensor, actuator)

Agent Programs

• A trivial agent program turns percept into action each time :

Simple Reflex Agents

• 下一個 action 只建立在 current percept, 而會無視以前的 percept history

• 只建立在 environment 是 full-observable

Model-based Reflex Agents

• Agent 會持續追蹤外部世界的變化

• 並更新內部的 internal state

Rule 跟 Simple Reflex agents 的是一樣的

Goal-based Reflex Agents

• 因為只有 current state of environment 沒辦法每次都能產出好的 action

• 所以基於 model 再新增 goal information

• Behavior 可以因 Goal 改變而被簡單改變

• Decision making 多了考慮未來的部分 "What/will"

Utility-based Reflex Agents

• 有 Goal 不見得能有 High quality behavior

• Utility-based agent 會挑選能夠 maximize expected utility 的 action

Learning Agents

• 使用 critic 的 feedback 來決定 performance element 要如何修改來加強 action

  • learning element => making improvements

  • performance element => selecting actions

• learn and improve in short run => better actions in long run

  • problem generator suggest new actions