Intelligent Agents Questions and Answers

Q: Why is a table driven agent impractical for most tasks?

The number of possible percept histories grows exponentially with the length of the agent's life, so even a chess playing table would need more entries than there are atoms in the observable universe. Lookup itself is fast, which is the only attraction of the design. Tables represent percept dependent behaviour perfectly well, and goal change is a separate objection. A design that is correct in principle and impossible in practice is still worth stating, because it defines what every real architecture is approximating.

Q: What does every practical agent architecture do that a lookup table does not?

The whole of agent design is finding compact procedures that reproduce the behaviour a vast table would have specified, and every architecture is a different way of doing so. Learning, internal state and lookahead are particular techniques rather than the general distinction, and simple reflex agents use none of them. Recognising the entire field of agent architecture as an exercise in compression makes the succession of designs easier to follow.

Q: What is an atomic state representation?

In an atomic representation states can only be compared for equality, so search algorithms can visit and remember them but cannot reason about what they have in common. It suffices for basic search and for Markov models. Factored representations use variables and structured ones use objects and relations. The choice of representation determines which algorithms are available, which is why it is settled before any algorithm is chosen.

Q: What does a factored representation allow that an atomic one does not?

Splitting a state into variables lets an algorithm notice that two states agree on most variables, which is what makes constraint propagation, Bayesian networks and planning heuristics possible. Objects and relations require the further step to structured representations. Memory efficiency is a side benefit and dynamism is unrelated. Each level of representation buys a class of inference that the level below cannot express, which is the reason the hierarchy is worth learning.

Q: Which representation is needed to say that one block is on another?

Expressing a relation between two objects requires a structured representation with objects, relations and the ability to quantify over them, which is what first order logic and relational planning languages provide. A factored representation can encode a fixed set of blocks clumsily but cannot generalise over them. Atomic representations cannot express it at all. The choice is not merely convenience: it determines what can be said as well as what can be computed.

Q: What is a percept in agent terminology?

A percept is the raw input at one moment, and the percept sequence is the whole history of them. Interpretation, action selection and internal state are things the agent does with percepts rather than percepts themselves. Being precise about the vocabulary matters because the agent function is defined as a mapping from percept sequences to actions, and a loose reading of either term makes that definition useless.

Q: Why is a vacuum cleaning agent a useful teaching example?

With two squares and two conditions, the whole environment has four states and the agent function can be written out as a table, so every architecture can be compared exactly. Real vacuum robots are far more complex, the performance measure is notoriously easy to get wrong, and internal state is needed as soon as the environment is made partially observable. A worked example small enough to enumerate is worth a great deal when the general case is intractable.

Q: What goes wrong if a cleaning agent is rewarded for dirt collected?

Rewarding the quantity collected rewards a cycle of dumping and recollecting, whereas rewarding a clean floor over time rewards the outcome actually wanted. It is the classic illustration of specifying the measure in terms of what is desired rather than what is easy to count. The other options describe behaviours that this measure does not encourage. An objective that admits a degenerate solution will find it, reliably, whenever the optimiser is any good.

Q: What is a semidynamic environment?

Chess with a clock is semidynamic: the board waits for the player but the performance measure penalises delay, so time matters even though the state does not change. A fully dynamic environment changes independently of the agent. The other options describe static, quasi-static and partially dynamic notions that are not the standard term. A category defined by whether the score rather than the state depends on time is a fine distinction that turns out to matter for real time design.

Q: What is an episodic environment?

In an episodic task the choice in one episode does not affect later ones, as in classifying items on a conveyor, so no lookahead is needed. In a sequential task, as in chess or driving, present choices have long term consequences and planning becomes essential. Resetting, discrete perception and end of run scoring are different notions. Whether decisions interact is one of the strongest determinants of how hard a task is.

Why is a table driven agent impractical for most tasks?

A.The table of percept sequences grows beyond any feasible size
B.Table lookup is slower than computing the action directly
C.Tables cannot represent actions that depend on the current percept
D.The table would have to be rewritten whenever the goal changed

Answer: A

The number of possible percept histories grows exponentially with the length of the agent's life, so even a chess playing table would need more entries than there are atoms in the observable universe. Lookup itself is fast, which is the only attraction of the design. Tables represent percept dependent behaviour perfectly well, and goal change is a separate objection. A design that is correct in principle and impossible in practice is still worth stating, because it defines what every real architecture is approximating.

Discuss in Forum

What does every practical agent architecture do that a lookup table does not?

A.Consider the consequences of actions before choosing between them
B.Compute the action from the percept rather than store every case
C.Learn from experience rather than following fixed behaviour
D.Maintain an internal model of the unobserved environment

Answer: B

The whole of agent design is finding compact procedures that reproduce the behaviour a vast table would have specified, and every architecture is a different way of doing so. Learning, internal state and lookahead are particular techniques rather than the general distinction, and simple reflex agents use none of them. Recognising the entire field of agent architecture as an exercise in compression makes the succession of designs easier to follow.

Discuss in Forum

What is an atomic state representation?

A.The state is a set of objects with relations between them
B.The state is a probability distribution over possible worlds
C.The state is a single indivisible token with no internal structure
D.The state is a vector of variables each with its own value

Answer: C

In an atomic representation states can only be compared for equality, so search algorithms can visit and remember them but cannot reason about what they have in common. It suffices for basic search and for Markov models. Factored representations use variables and structured ones use objects and relations. The choice of representation determines which algorithms are available, which is why it is settled before any algorithm is chosen.

Discuss in Forum

What does a factored representation allow that an atomic one does not?

A.Storing more states in the same amount of memory
B.Representing objects and the relations between them
C.Handling environments that change while the agent thinks
D.Reasoning about states that share some variable values

Answer: D

Splitting a state into variables lets an algorithm notice that two states agree on most variables, which is what makes constraint propagation, Bayesian networks and planning heuristics possible. Objects and relations require the further step to structured representations. Memory efficiency is a side benefit and dynamism is unrelated. Each level of representation buys a class of inference that the level below cannot express, which is the reason the hierarchy is worth learning.

Discuss in Forum

Which representation is needed to say that one block is on another?

A.Structured
B.Atomic, since the whole configuration is a single state
C.Factored representation
D.Any of them, since the choice is a matter of convenience

Answer: A

Expressing a relation between two objects requires a structured representation with objects, relations and the ability to quantify over them, which is what first order logic and relational planning languages provide. A factored representation can encode a fixed set of blocks clumsily but cannot generalise over them. Atomic representations cannot express it at all. The choice is not merely convenience: it determines what can be said as well as what can be computed.

Discuss in Forum

What is a percept in agent terminology?

A.The internal state the agent maintains between inputs
B.The agent's perceptual input at a given instant
C.The agent's interpretation of what it has sensed
D.The action the agent selects in response to its input

Answer: B

A percept is the raw input at one moment, and the percept sequence is the whole history of them. Interpretation, action selection and internal state are things the agent does with percepts rather than percepts themselves. Being precise about the vocabulary matters because the agent function is defined as a mapping from percept sequences to actions, and a loose reading of either term makes that definition useless.

Discuss in Forum

Why is a vacuum cleaning agent a useful teaching example?

A.Its performance measure is impossible to specify incorrectly
B.It requires no internal state under any circumstances
C.Its environment is small enough to enumerate completely
D.It is the simplest agent that can be built with real hardware

Answer: C

With two squares and two conditions, the whole environment has four states and the agent function can be written out as a table, so every architecture can be compared exactly. Real vacuum robots are far more complex, the performance measure is notoriously easy to get wrong, and internal state is needed as soon as the environment is made partially observable. A worked example small enough to enumerate is worth a great deal when the general case is intractable.

Discuss in Forum

What goes wrong if a cleaning agent is rewarded for dirt collected?

A.It will clean too slowly to be useful in practice
B.It will ignore squares that are only lightly soiled
C.It will refuse to move between squares at all
D.It may dump dirt and collect it again to raise its score

Answer: D

Rewarding the quantity collected rewards a cycle of dumping and recollecting, whereas rewarding a clean floor over time rewards the outcome actually wanted. It is the classic illustration of specifying the measure in terms of what is desired rather than what is easy to count. The other options describe behaviours that this measure does not encourage. An objective that admits a degenerate solution will find it, reliably, whenever the optimiser is any good.

Discuss in Forum

What is a semidynamic environment?

A.The world does not change but the agent's score does with time
B.The world changes only when the agent takes an action
C.The world changes slowly enough for the agent to keep up
D.Only part of the world changes while the agent deliberates

Answer: A

Chess with a clock is semidynamic: the board waits for the player but the performance measure penalises delay, so time matters even though the state does not change. A fully dynamic environment changes independently of the agent. The other options describe static, quasi-static and partially dynamic notions that are not the standard term. A category defined by whether the score rather than the state depends on time is a fine distinction that turns out to matter for real time design.

Discuss in Forum

10.

What is an episodic environment?

A.The agent's performance is measured once at the end of a long run
B.Each decision is independent of those before and after it
C.The environment resets to its initial state after each action
D.The agent receives its percepts in discrete episodes rather than continuously

Answer: B

In an episodic task the choice in one episode does not affect later ones, as in classifying items on a conveyor, so no lookahead is needed. In a sequential task, as in chess or driving, present choices have long term consequences and planning becomes essential. Resetting, discrete perception and end of run scoring are different notions. Whether decisions interact is one of the strongest determinants of how hard a task is.

Discuss in Forum

Loading…

Topics

Other Subjects

Tags: Intelligent Agents Questions and Answers || Intelligent Agents MCQ Questions and Answers || Intelligent Agents GK Questions and Answers || Intelligent Agents GK MCQ Questions || Intelligent Agents Multiple Choice Questions and Answers || Intelligent Agents GK || GK on Intelligent Agents || Artificial Intelligence Questions and Answers || Artificial Intelligence MCQ Questions and Answers || Artificial Intelligence GK Questions and Answers || GK on Artificial Intelligence

Intelligent Agents Questions and Answers

Topics

Other Subjects

About Us

Stay informed about all important Dates & Notifications