Agents

Question 1

Describe the diagram on the agent-oriented perspective on AI

Answer 1

Look in notes if you don’t remember :p

Question 2

____________ are things that could change the state of an environment

Answer 2

Actuators

Question 3

Sensors process _________ from the environment

Answer 3

percepts

Question 4

What is a percept?

Answer 4

An agent’s perceptual input at any point in time

Question 5

And agent’s percept _________ is the order of everything it has perceived

Answer 5

sequence

Question 6

In general, an agent’s choice of actions can depend on its entire ______ _______

Answer 6

percept history

Question 7

A performance measure measures the degree of the “______ _____” the agent does

Answer 7

right thing

Question 8

What are 4 things a agent needs to be rational?

Answer 8

• A performance measure that defines its criterion for success
• Some prior knowledge of the environment
• To know the actions it can perform
• Its percept sequence to date

Question 9

What is the Definition of a rational agent? (Memorize this)

Answer 9

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.

Question 10

To tell if an agent is ________, we have to take into account the relevant performance measure

Answer 10

rational

Question 11

An agent might do an ________ _______ actions to learn more about the environment

Answer 11

information gathering

Question 12

What does it mean for an agent to be autonomous?

Answer 12

They can operate on their own without a lot of prior information from their designers

Question 13

We expect rational agents to deal with ________ and _____ information, and to learn new _______ when necessary

Answer 13

impartial, unknown, behaviours

Question 14

What does the PEAS model stand for?

Answer 14

• Performance measure
• Environment
• Actuators
• Sensors

Question 15

Describe a taxi-driving agent using the PEAS model

Answer 15

performance measure: safety, speed, legality, passenger comfort, profit

environment: roads, including other traffic, pedestrians, weather conditions
actuators: steering, breaks, accelerator, turn signals, horn, lights
sensors: camera, sonar, speedometer, odometer, GPS, engine sensors

Question 16

What are the 5 property pairs of environments?

Answer 16

• Fully Observable vs Partially Observable
• Deterministic vs Stochastic
• Static vs Dynamic
• Discrete vs Continuous
• Known vs Unknown

Question 17

When is an environment Fully Observable?

Answer 17

If an agents sensors give it access to the complete environment at each point in time

Question 18

When is an environment Partially Observable?

Answer 18

only a part of the environment can be sensed at each point in time

Question 19

When is an environment Deterministic?

Answer 19

if the next state of the environment is completely determined by the previous state sequence, and the action of the agent

Question 20

When is an environment Stochastic?

Answer 20

if the next state of the environment is NOT completely determined by the previous state sequence, and the action of the agent

Question 21

When is an environment Static?

Answer 21

if the environment cannot change while the agent is thinking

Question 22

When is an environment Dynamic?

Answer 22

if the environment can change while the agent is thinking

Question 23

When is an environment Discrete?

Answer 23

if time is handled in discrete sequential steps (like the discrete ticking of a CPU)

Question 24

When is an environment Continuous?

Answer 24

time is handled as a continuous stream (as in physics)

Question 25

When is an environment known?

Answer 25

if the outcomes of all actions are known ahead of time by the agent

Question 26

When is an environment unknown?

Answer 26

the outcomes of some actions are not known ahead of time, meaning the agent must test things out to find out what will happen

Question 27

What are 6 different structures of agents?

Answer 27

• Table-driven
• Simple Reflex
• Model-based Reflex
• Goal-based
• Utility
• Learning

Question 28

A table-driven agent uses a table of actions indexed by ______-________

Answer 28

percept sequences

Question 29

Using a table-driven agent is unreasonable in even _________ complicated situations

Answer 29

moderately

Question 30

A ________ _______ agent decides what to do based solely on the current percept, and ignores all previous percepts

Answer 30

simple reflex

Question 31

Simple reflex agents consist of _______-______ rules

Answer 31

condition action

Question 32

What are 2 cons of the simple reflex structure?

Answer 32

• Single percepts are not enough to determine the right thing
• Can get stuck in infinite loops

Question 33

Model-based reflex agents save information about the external world using _______ state

Answer 33

internal

Question 34

A __________ agent can simulate actions with their internal model

Answer 34

Model-based

Question 35

What are goals?

Answer 35

A description of desirable states

Question 36

Utility agents are ______-______ agents that also take into account the quality of the _____ sequences used to achieve the goal

Answer 36

goal-based, action

Question 37

A ______ function measures the agent’s “happiness” with a set of actions

Answer 37

utility

Question 38

Utility-maximizing agents are the goal of much ____ research

Answer 38

AI

Question 39

A _______ agent can modify its components to improve overall performance

Answer 39

learning