CS 188 Spring: Exam Prep 2 Solutions
for AI Concepts
Rational Agent - correct answersThe entity that has goals/preferences and tries to perform a series of
actions that yield the best-expected outcome
Environment - correct answersWhere rational agents exist, specific to given instantiation of agent
World - correct answersEnvironment + agents
Reflex agent - correct answersDoesn't think about consequences of its actions, selects on action based
solely on current state of the world
Planning Agents - correct answersMaintain a model of the world to simulate performing various actions
PEAS - correct answersPerformance, Measure, Environment, Actuators, Sensors
Design - correct answersOf an agent depends heavily on type of environment the agents acts upon.
Characterized by partially observable, stochastic, multi-agent, dynamic, static, known physics, etc.
Partially Observable - correct answersAgent doesn't have full information on state, agent must have an
internal estimate of the state of the world
Fully Observable - correct answersAgent has full information on the state
Stochastic - correct answersUncertainty in the transition model, taking action in a specific state may
have multiple possible outcomes w/ different probabilities (unpredictable multiple possible outcomes)
Deterministic - correct answersTaking an action in a state has a single outcome that is guaranteed to
happen
Multi-Agent - correct answersAgents acts in the environment along w/ other agents
Agent needs to randomize to avoid predictability
Dynamic - correct answersChange as the environment interacts with it
Known Physics - correct answersTransition model is known to agent, can use when planning a path
Unknown Physics - correct answersAgent need to take actions deliberately to learn unknown dynamics
Search Problem - correct answersGiven our agent's current state (configuration w/in its environment)
how can we arrive at a new state that satisfies this goals in the best possible way?
State space, actions transition model, action cost, start state, goal test
Plan - correct answersPath from the start to goal
for AI Concepts
Rational Agent - correct answersThe entity that has goals/preferences and tries to perform a series of
actions that yield the best-expected outcome
Environment - correct answersWhere rational agents exist, specific to given instantiation of agent
World - correct answersEnvironment + agents
Reflex agent - correct answersDoesn't think about consequences of its actions, selects on action based
solely on current state of the world
Planning Agents - correct answersMaintain a model of the world to simulate performing various actions
PEAS - correct answersPerformance, Measure, Environment, Actuators, Sensors
Design - correct answersOf an agent depends heavily on type of environment the agents acts upon.
Characterized by partially observable, stochastic, multi-agent, dynamic, static, known physics, etc.
Partially Observable - correct answersAgent doesn't have full information on state, agent must have an
internal estimate of the state of the world
Fully Observable - correct answersAgent has full information on the state
Stochastic - correct answersUncertainty in the transition model, taking action in a specific state may
have multiple possible outcomes w/ different probabilities (unpredictable multiple possible outcomes)
Deterministic - correct answersTaking an action in a state has a single outcome that is guaranteed to
happen
Multi-Agent - correct answersAgents acts in the environment along w/ other agents
Agent needs to randomize to avoid predictability
Dynamic - correct answersChange as the environment interacts with it
Known Physics - correct answersTransition model is known to agent, can use when planning a path
Unknown Physics - correct answersAgent need to take actions deliberately to learn unknown dynamics
Search Problem - correct answersGiven our agent's current state (configuration w/in its environment)
how can we arrive at a new state that satisfies this goals in the best possible way?
State space, actions transition model, action cost, start state, goal test
Plan - correct answersPath from the start to goal
