Russell and Norvig, Chapter 2: Intelligent Agents

Intelligent Agents
  AGENT - perceives through sensors and acts in an environment through
          effectors. 
    - [Figure 2.1]
    - percepts (sensor readings)
      - percept sequence is everything perceived so far
      - sounds, images, text, facts, rules, knowledge
    - actions
      -  manipulation of effectors that affect the environment
    - goals: performance measure
      - more effective, more efficient
    - environment
      - accessible vs. inaccessible
        - sensors detect all aspects of environment relative to choosing an
          action 
      - deterministic vs. nondeterministic
        - is the next state entirely determined by the previous state and
          action 
      - episodic vs. nonepisodic
        - episodes do not depend on actions in previous episodes
      - static vs. dynamic
      - discrete vs. continuous
      - [Figure 2.15]
    - agent program
      - [Figure 2.4]
      - background knowledge
    - example: TAGER student
      - percepts: sounds, images, text
      - actions: questions (class/email), homeworks, exams
      - goals: score high, brown-nose teacher, ask questions, 
      - environment: ~accessible, ~deterministic, nonepisodic, dynamic,
                     ~discrete
      - agent program
        - lots of background knowledge, infant.zip, kinder.zip - cse5303.zip
        - knowledge representation and reasoning
        - search
        - planning
        - natural language processing
        - uncertainty reasoning
        - learning
        - vision and speech processing
  - RATIONAL AGENT
    - for each percept sequence, agent choses an action to maximize its
      performance measure based on the percept sequence and any built-in
      knowledge 
  - truly intelligent TAGER student: try Turing test
  - agent types
    - reflex agent
      - rule-based
      - e.g. factorial agent
    - memory-based agents
      - rule-based plus learned rules
      - e.g., factorial agent that remembers previous solutions
    - goal-based agents
      - planner
      - e,g, route planner: getting from Dallas to UTA
    - utility-based agents
      - maximize continuous-valued goals
      - e.g., planner: tradeoff between homework score and sleep