1. Dr. Alaa Sagheer 2010-2011

2. Chapter 2 Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

3.  Agents and environments  Rationality  PEAS (Performance measure, Environment, Actuators, Sensors)  Environments types  Agents types Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

4.  An agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators  Human agent: eyes, ears, and other organs for sensors;  hands, legs, mouth, and other body parts for actuators Robotic agent: cameras and infrared range finders for sensors;  various motors for actuators Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

5.  The agent function maps from percept histories to actions: [f: P*  A ] The agent program runs on the physical architecture to produce f agent = architecture + program Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

6.  Environment: square A and B  Percepts: [location and content] e.g. [A, Dirty]  Actions: left, right, suck, and no-op Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

7. Percept sequenceAction [A,Clean]Right [A, Dirty]Suck [B, Clean]Left [B, Dirty]Suck [A, Clean],[A, Clean]Right [A, Clean],[A, Dirty]Suck …… Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

8.  function REFLEX-VACUUM-AGENT ([location, status]) return an action if status == Dirty then return Suck else if location == A then return Right else if location == B then return Left  What is the right function? Can it be implemented in a small agent program? Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

9.  An agent should strive to "do the right thing", based on what it can perceive and the actions it can perform. The right action is the one that will cause the agent to be most successful  Performance measure: An objective criterion for success of an agent's behavior  E.g., performance measure of a vacuum- cleaner agent could be amount of dirt cleaned up, amount of time taken, amount of electricity consumed, amount of noise generated, etc. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

10.  Rational Agent: For each possible percept sequence, a rational agent should select an action that is expected to maximize its performance measure, given the evidence provided by the percept sequence and whatever built-in knowledge the agent has. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

11.  Rationality is distinct from omniscience (all-knowing with infinite knowledge)  Agents can perform actions in order to modify future percepts so as to obtain useful information (information gathering, exploration)  An agent is autonomous if its behavior is determined by its own experience (with ability to learn and adapt) Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

12.  A rational agent is one that does the right thing › Every entry in the table is filled out correctly  What is the right thing? › Approximation: the most successful agent › Measure of success?  Performance measure should be objective › E.g. the amount of dirt cleaned within a certain time › E.g. how clean the floor is ›…›…  Better to design performance measures according to what is wanted in the environment instead of how the agents should behave. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

13.  What is rational at a given time depends on four things: › Performance measure › Prior environment knowledge › Actions › Percept sequence to date (sensors)  A rational agent chooses whichever action maximizes the expected value of the performance measure given the percept sequence to date and prior environment knowledge Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

14.  Rationality  omniscience › An omniscient agent knows the actual outcome of its actions.  Rationality  perfection › Rationality maximizes expected performance, while perfection maximizes actual performance. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

15.  The proposed definition requires: › Information gathering/exploration  To maximize future rewards › Learn from percepts  Extending prior knowledge › Agent autonomy  Compensate for incorrect prior knowledge Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

16.  Task environment › a problem for which a rational agents is the solution › specified by the PEAS description:  Performance measure, Environment, Actuators, Sensors  Must first specify the setting for intelligent agent design  Consider, e.g., the task of designing an automated taxi: › Performance measure › Environment › Actuators › Sensors Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

17.  Must first specify the setting for intelligent agent design  Consider, e.g., the task of designing an automated taxi driver: › Performance measure: Safe, fast, legal, comfortable trip, maximize profits › Environment: Roads, other traffic, pedestrians, customers › Actuators: Steering wheel, accelerator, brake, signal, horn › Sensors: Cameras, sonar, speedometer, GPS, odometer, engine sensors, keyboard Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

18.  Agent: Medical diagnosis system  Performance measure: Healthy patient, minimize costs, lawsuits  Environment: Patient, hospital, staff  Actuators: Screen display (questions, tests, diagnoses, treatments, referrals)  Sensors: Keyboard (entry of symptoms, findings, patient's answers) Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

19.  Agent: Part-picking robot  Performance measure: Percentage of parts in correct bins  Environment: Conveyor belt with parts, bins  Actuators: Jointed arm and hand  Sensors: Camera, joint angle sensors Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

20.  Agent: Interactive English tutor  Performance measure: Maximize student's score on test  Environment: Set of students  Actuators: Screen display (exercises, suggestions, corrections)  Sensors: Keyboard Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

21.  Fully observable (vs. partially observable): An agent's sensors give it access to the complete state of the environment at each point in time.  Deterministic (vs. stochastic): The next state of the environment is completely determined by the current state and the action executed by the agent. (If the environment is deterministic except for the actions of other agents, then the environment is strategic)  Episodic (vs. sequential): The agent's experience is divided into atomic "episodes" (each episode consists of the agent perceiving and then performing a single action), and the choice of action in each episode depends only on the episode itself. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

22.  Static (vs. dynamic): The environment is unchanged while an agent is deliberating. (The environment is semi dynamic if the environment itself does not change with the passage of time but the agent's performance score does)  Discrete (vs. continuous): A limited number of distinct, clearly defined percepts and actions.  Single agent (vs. multiagent): An agent operating by itself in an environment. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

23. SolitaireBackgammomIntenet shoppingTaxi Observable?? Deterministic?? Episodic?? Static?? Discrete?? Single-agent?? Fully vs. partially observable: an environment is full observable when the sensors can detect all aspects that are relevant to the choice of action. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

24. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic?? Episodic?? Static?? Discrete?? Single-agent?? Fully vs. partially observable: an environment is full observable when the sensors can detect all aspects that are relevant to the choice of action. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

25. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic?? Episodic?? Static?? Discrete?? Single-agent?? Deterministic vs. stochastic: if the next environment state is completely determined by the current state the executed action then the environment is deterministic. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

26. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic?? Static?? Discrete?? Single-agent?? Deterministic vs. stochastic: if the next environment state is completely determined by the current state the executed action then the environment is deterministic. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

27. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic?? Static?? Discrete?? Single-agent?? Episodic vs. sequential : In an episodic environment the agent’s experience can be divided into atomic steps where the agents perceives and then performs A single action. The choice of action depends only on the episode itself Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

28. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static?? Discrete?? Single-agent?? Episodic vs. sequential: In an episodic environment the agent’s experience can be divided into atomic steps where the agents perceives and then performs A single action. The choice of action depends only on the episode itself Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

29. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static?? Discrete?? Single-agent?? Static vs. dynamic: If the environment can change while the agent is choosing an action, the environment is dynamic. Semi- dynamic if the agent’s performance changes even when the environment remains the same. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

30. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static??YES SEMINO Discrete?? Single-agent?? Static vs. dynamic: If the environment can change while the agent is choosing an action, the environment is dynamic. Semi- dynamic if the agent’s performance changes even when the environment remains the same. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

31. SolitaireBackgammomInternet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static??YES SEMINO Discrete?? Single-agent?? Discrete vs. continuous: This distinction can be applied to the state of the environment, the way time is handled and to the percepts/actions of the agent. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

32. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static??YES SEMINO Discrete??YES NO Single-agent?? Discrete vs. continuous: This distinction can be applied to the state of the environment, the way time is handled and to the percepts/actions of the agent. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

33. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static??YES SEMINO Discrete??YES NO Single-agent?? Single vs. multi-agent: Does the environment contain other agents who are also maximizing some performance measure that depends on the current agent’s actions? Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

34. SolitaireBackgammomIntenet shoppingTaxi Observable??FULL PARTIAL Deterministic??YESNOYESNO Episodic??NO Static??YES SEMINO Discrete??YES NO Single-agent??YESNO Single vs. multi-agent: Does the environment contain other agents who are also maximizing some performance measure that depends on the current agent’s actions? Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

35.  The simplest environment is › Fully observable, deterministic, episodic, static, discrete and single-agent.  Most real situations are: › Partially observable, stochastic, sequential, dynamic, continuous and multi-agent. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

36.  An agent is completely specified by the agent function mapping percept sequences to actions  One agent function (or a small equivalence class) is rational  Aim: find a way to implement the rational agent function concisely Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

37.  Function TABLE-DRIVEN_AGENT(percept) returns an action  static : percepts, a sequence initially empty table, a table of actions, indexed by percept sequence append percept to the end of percepts action  LOOKUP(percepts, table) return action Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

38.  Four basic types in order of increasing generality:  Simple reflex agents  Model-based reflex agents  Goal-based agents  Utility-based agents All these can be turned into learning agents. Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

39.  Select action on the basis of only the current percept ignoring the rest of history. › E.g. the vacuum-agent  Large reduction in possible percept/action situations (example on next page)  Implemented through condition-action rules › If dirty then suck Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

41. function SIMPLE-REFLEX-AGENT(percept) returns an action static : rules, a set of condition-action rules state  INTERPRET-INPUT(percept) rule  RULE-MATCH(state, rule) action  RULE-ACTION[rule] return action Will only work if the environment is fully observable  E.g., determination of braking in cars without centrally mounted brake lights Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

42.  To tackle partially observable environments › Maintain internal state  Over time update state using world knowledge › How does the world change. › How do actions affect world.  Model of World Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

44.  The agent needs a goal to know which situations are desirable › Things become difficult when long sequences of actions are required to find the goal  Typically investigated in search and planning research  Major difference: future is taken into account  Is more flexible since knowledge is represented explicitly and can be manipulated Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

45.  Certain goals can be reached in different ways › Some are better, have a higher utility  Utility function maps a (sequence of) state(s) onto a real number.  Improves on goals: › Selecting between conflicting goals › Select appropriately between several goals based on likelihood of success Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer

46. AAll previous agent programs describe methods for selecting actions ›Y›Yet the origin of these programs is not provided ›L›Learning mechanisms can be used to perform this task ›T›Teach them instead of instructing them ›A›Advantage is the robustness of the program toward initially unknown environments

47.  Learning element: introduce improvements in performance element. › Critic provides feedback on agents performance based on fixed performance standard  Performance element: selecting actions based on percepts › Corresponds to the previous agent programs  Problem generator: suggests actions that will lead to new and informative experiences › Exploration Artificial Intelligence Ch2: Intelligent Agents Dr. Alaa Sagheer