|
Created by August Edström
over 6 years ago
|
|
Question | Answer |
What are the most important aspects of agents? – According to Russel & Norvig | Percieving its environment and acting upon that environment |
Name two reasons why we would want to adopt the multiagent paradigm | Omni-present, connected, human-oriented software |
What is the difference between an agent and an object? | Agents are autonomous, Smart, flexible behaviour, Active (an agents has at least one thread in which it has active control) |
Explain the intentional stance by means of an example | A bird will fly away because it knows the cat is coming and is afraid of getting eaten |
Specify PEAS for the IBM Watson Jeopardy system | PEAS are specifications in designing a rational agent. P - right answer, winning the game, E - Jeoperdy game, A - Text, S - keyboard entry? |
Indicate for a self-driving car what the environment is (observable, agents, deterministic, episodic, static, concrete) | Partially, Multi, Stochastic, Sequential, Dynamic, Continuous |
What is the difference between a model-based and a simple reflex agent? | Model-based cares about history, simple reflex does not |
Explain the architecture of a utility-based agent | It is goal-oriented but can achieve the goal in a number of ways. Utility is determined by the user |
What is a dialogue protocol? Explain by giving a concrete example of a protocol | A schema of how a dialogue should be conducted. Agent A sends Request to agent B. B Agrees. B Inform A. |
Name three reasons why we would want to define and study formal dialogue protocols | A shared understanding of the meaning of utterances, sequences of utterances and dialogues for agents To enable different agent languages and protocols o be compared To enable agent languages and protocols to be readily implemented in production systems |
Take the payoff matrix as given on slide 52 (prisoner’s dilemma) – What is the dominant strategy for a1? And for a2? – Which outcome(s) has (have) the highest social welfare? – Which outcome(s) is (are) pareto optimal? | - To confess -Not confess -Not confess |
What is the definition of a Nash equilibrium? – Does the prisoner’s dilemma (slide 54) have a Nash equilibrium? | That no individual has the incentive to change strategy without an agreement beforehand -Yes, confess. |
Want to create your own Flashcards for free with GoConqr? Learn more.