Foundations of Artificial Intelligence February 19, 2020 — 3. Introduction: Rational Agents Foundations of Artificial Intelligence 3. Introduction: Rational Agents 3.1 Agents Malte Helmert and Thomas Keller 3.2 Rationality University of Basel 3.3 Summary February 19, 2020 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 1 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 2 / 19 3. Introduction: Rational Agents Agents Introduction: Overview Chapter overview: introduction 3.1 Agents ◮ 1. What is Artificial Intelligence? ◮ 2. AI Past and Present ◮ 3. Rational Agents ◮ 4. Environments and Problem Solving Methods M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 3 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 4 / 19
3. Introduction: Rational Agents Agents 3. Introduction: Rational Agents Agents Heterogeneous Application Areas Agents AI systems are used for very different tasks: sensors ◮ controlling manufacturing plants percepts ◮ detecting spam emails ? environment agent ◮ intra-logistic systems in warehouses actions ◮ giving shopping advice on the Internet actuators ◮ playing board games ◮ finding faults in logic circuits Agents ◮ . . . ◮ agent functions map sequences of observations to actions: f : P + → A How do we capture this diversity in a systematic framework emphasizing commonalities and differences? ◮ agent program: runs on physical architecture and computes f common metaphor: rational agents and their environments Examples: human, robot, web crawler, thermostat, OS scheduler German: rationale Agenten, Umgebungen German: Agenten, Agentenfunktion, Wahrnehmung, Aktion M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 5 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 6 / 19 3. Introduction: Rational Agents Agents 3. Introduction: Rational Agents Agents Introducing: an Agent Vacuum Domain A B ◮ observations: location and cleanness of current room: � a , clean � , � a , dirty � , � b , clean � , � b , dirty � ◮ actions: left, right, suck, wait M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 7 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 8 / 19
3. Introduction: Rational Agents Agents 3. Introduction: Rational Agents Agents Vacuum Agent Reflexive Agents Reflexive agents compute next action only based on last observation in sequence: ◮ very simple model a possible agent function: ◮ very restricted observation sequence action ◮ corresponds to Mealy automaton (a kind of DFA) � a , clean � right with only 1 state � a , dirty � suck � b , clean � left ◮ practical examples? � b , dirty � suck German: reflexiver Agent � a , clean � , � b , clean � left � a , clean � , � b , dirty � suck Example (A Reflexive Vacuum Agent) . . . . . . def reflex-vacuum-agent( location , status ): if status = dirty: return suck else if location = a: return right else if location = b: return left M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 9 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 10 / 19 3. Introduction: Rational Agents Agents 3. Introduction: Rational Agents Rationality Evaluating Agent Functions 3.2 Rationality What is the right agent function? M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 11 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 12 / 19
3. Introduction: Rational Agents Rationality 3. Introduction: Rational Agents Rationality Rationality Is Our Agent Perfectly Rational? Rational Behavior Question: Is the reflexive vacuum agent Evaluate behavior of agents with performance measure of the example perfectly rational? (related terms: utility, cost). depends on performance measure and environment! perfect rationality: ◮ Do actions reliably have the desired effect? ◮ always select an action maximizing ◮ Do we know the initial situation? ◮ expected value of future performance ◮ Can new dirt be produced while the agent is acting? ◮ given available information (observations so far) German: Performance-Mass, Nutzen, Kosten, perfekte Rationalit¨ at M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 13 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 14 / 19 3. Introduction: Rational Agents Rationality 3. Introduction: Rational Agents Rationality Rational Vacuum Agent Rationality: Discussion Example (Vacuum Agent) performance measure: ◮ perfect rationality � = omniscience ◮ +100 units for each cleaned cell ◮ incomplete information (due to limited observations) reduces achievable utility ◮ − 10 units for each suck action ◮ perfect rationality � = perfect prediction of future ◮ − 1 units for each left / right action ◮ uncertain behavior of environment (e.g., stochastic action effects) reduces achievable utility environment: ◮ perfect rationality is rarely achievable ◮ actions and observations reliable ◮ limited computational power � bounded rationality ◮ world only changes through actions of the agent German: begrenzte Rationalit¨ at ◮ all initial situations equally probable How should a perfect agent behave? M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 15 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 16 / 19
3. Introduction: Rational Agents Summary 3. Introduction: Rational Agents Summary Summary (1) common metaphor for AI systems: rational agents 3.3 Summary agent interacts with environment: ◮ sensors perceive observations about state of the environment ◮ actuators perform actions modifying the environment ◮ formally: agent function maps observation sequences to actions ◮ reflexive agent: agent function only based on last observation M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 17 / 19 M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 18 / 19 3. Introduction: Rational Agents Summary Summary (2) rational agents: ◮ try to maximize performance measure (utility) ◮ perfect rationality: achieve maximal utility in expectation given available information ◮ for “interesting” problems rarely achievable � bounded rationality M. Helmert, T. Keller (University of Basel) Foundations of Artificial Intelligence February 19, 2020 19 / 19
Recommend
More recommend
