Artificial Intelligence - Agents Overview

Artificial Intelligence (AI) - Agent

• An agent is anything that senses the environment using sensors and acts upon that environment through its actuators.
• A sensor receives input from the environment.
• An actuator helps interact with the environment.
• Percept is the perceptual input of an agent at a given moment.
• Percept sequence is the complete history of everything the agent has perceived.
• Agent function (control policy function) is a set of condition-rules based on which an agent can make decision actions.

Applications of Agents

• AI is successfully used in:
  • Finance
  • Robotics
  • Games
  • Medicine
  • The web

Vacuum Cleaner Agent

• Example: A vacuum cleaner agent's percepts might be location and status (e.g., [A, Dirty]).
• Actions: Left, Right, Suck, NoOp.
• Function Vacuum-Agent([location, status]) returns an action.
  • If status is Dirty, then return Suck.
  • If location is A, then return Right.
  • If location is B, then return Left.

Rational Agent

• Performance measure is an objective criterion (utility function) for assessing an agent's success.
• Performance measures for a vacuum-cleaner agent include amount of dirt cleaned, time taken, electricity consumed, and noise level.
• Self-driving car performance measures include time to destination (minimized), safety, predictability to other agents, and reliability.
• Game-playing agent performance measures include win/loss percentage (maximized), robustness, and unpredictability (to confuse the opponent).
• A rational agent, for each possible percept sequence, selects an action that maximizes its performance measure given the percept sequence evidence. This includes the agent's built-in knowledge.
• Rationality means doing the right thing.
• Rationality involves maximizing expected performance, information gathering, exploration, and learning.

Variables in Python

• Python is case-sensitive.
• Types of variables include string, float, integer, Boolean, complex, list, tuple, set, and dictionary.
• Number variables (integer, float):
  • To define: x = 5, y = 5.5.
  • Conversion between data types: x1 = str(x), y2 = bool(y), x2 = complex(x, y).
  • Arithmetic operations: print(x + y, x - y, x * y, x / y, x**y, x % y, x // y).
• String variables:
  • To define: name = "ali".
  • Accessing characters: print(a[5]), print(a[4:12]).
  • Repeating strings/characters: print(a*3).
  • Splitting strings: print(a.split()).
  • Partitioning strings: print(a.partition('')).
  • Finding characters/words: print(a.find("artificial")).
  • Replacing characters: print(a.replace('a', 'z')).
  • Counting characters/words: print(a.count("e")).
  • Converting case: print(a.upper()), print(a.lower()), print(a.capitalize()), print(a.title()), print(a.swapcase()).
• Boolean variables:
  • To define: a = True, x = False.