Notes
Mark Chen |
28 Feb 2021
报告错误
如果你发现该网页中存在错误/显示异常，可以从以下两种方式向我们报告错误，我们会尽快修复：
- 使用 CS Club 网站错误 为主题，附上错误截图或描述及网址后发送邮件到 286988023@qq.com
- 在我们的网站代码仓库中创建一个 issue 并在 issue 中描述问题 点击链接前往Github仓库
Chapter 2 Intelligent Agents
In chapter 1, we have mentioned the concept of rational agents as the central approach to artificial intelligence, where rational agents are systems that can be reasonably be called intelligent.
2.1 Agents and Environments
agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators.
We use Percept to refer to the agent’s perceptual inputs at any given instant, and a program’s Percept Sequence is the complete history of everything the agent has perceived.
An agent’s choice of action at any given instant can depend on the entire percept sequence observed to date, but not on anything it hasn’t perceived.
We say that the agent’s behavior is described by the agent function that maps any given percept sequence to an action.
\[Percepts \rightarrow Agent\space Function(Percepts) \rightarrow Actions\]Agent function for an artificial agent will be implemented by an agent program. It should be notice that the agent function is an abstract mathematical description, the agent program is a concrete implementation, running within some physical system.
Model’s building up: Let’s suppose a space called the ‘vacuum-cleaner world’, shown in the figure below
There are two areas, area A and B, where some dirt will appear in both rooms. The vacuum cleaner can both ‘Suck’, ‘Go Right’ and ‘Go Left’.
Artificial Intelligence operates at where the artifacts have significant computational resources and the task environment requires nontrival decision making.
2.2 Good Behavior: The Concept of Rationality
A rational agent is one that does the right thing. We can define the standard of ‘doing right thing’ by considering the agent’s behavior. If the sequence is desirable, then the agent has performed well.
Performance Measure is one process that evaluate whether a given sequence of environment state sequence can be seen as ‘desirable’. Notice that we are talking about the Environment Statas, not the Agent Statas
As a general rule, it is better to design performance measures according to what one actually wants in the environment, instead of thinking about the behavior of rational agent.
