Deck 9: Machines That Can Learn

Question 1

Neural computing involves processing information by means of changing states of networks formed by interconnecting extremely large numbers of simple processing elements that interact with one another by exchanging signals.

Answer 1

True

Question 2

Fuzzy logic allows for the partial description of a rule.

Answer 2

True

Question 3

Fuzzy logic is key to developing computers that can think.

Answer 3

True

Question 4

The human language has evolved to allow for the conveyance of meaning through semantic approximation rather than precise content

Question 5

Fuzzy logic deals with the likelihood that something has a particular property while probability deals with the degree to which a particular property is present in something

Question 6

Fuzzy logic is a new development in the computing industry that emerged as a result of the development of Internet search engines.

Question 7

Fuzzy logic can deal with any degree of change in input.

Question 8

When continuous variables are involved, traditional rule-based systems tend to provide more accurate results than fuzzy logic.

Question 9

Fuzzy logic allows for the modeling of contradiction within a knowledge base.

Question 10

One of the limitations of a fuzzy logic system is that each rule is dependant upon all of the other rules in the knowledge base.

Question 11

Under highly complex situations, fuzzy logic rules are highly verifiable.

Question 12

Artificial neural networks are simple computer-based programs whose primary function is to construct models of a problem space based upon trial and error.

Question 13

The transmission of signals from one neuron to another in an artificial neural network occurs at the neural synapse via a complex chemical process in which specific substances, called neurotransmitters, are released from one side of the synaptic junction.

Question 14

The ANN involves an interconnected system of nodes called neurodes that are associated with one or more weighted connections that are equivalent to human neural synapses, inside the memory of a digital computer. The construction of the ANN is one of multiple layers with the connections running between the layers.

Question 15

The basic structure a typical neurode consists of is a set of weighted input connections, a bias input, a state function, a nonlinear transfer function, and an output connection.

Question 16

The purpose of a state function is to consolidate the weights of the various inputs to the neurode into a single value that can be passed to the transfer function for processing.

Question 17

The primary purpose of the learning paradigm is to serve as the vehicle by which the summed information is passed on as output.

Question 18

An advantage of automated neural nets is the elimination of the need for direct input from experts.

Question 19

The GA's smallest unit is called a chromosome. The chromosome represents the smallest unit of information in the problem domain and can be thought of as the basic building block for a possible solution.

Question 20

The GA's smallest unit is called a gene. The gene represents the smallest unit of information in the problem domain and can be thought of as the basic building block for a possible solution.

Question 21

______________ is a method of reasoning that allows for the partial description of a rule. Combining this approach to reasoning with the realm of digital processors has results in a class of computer applications that can "learn" from their mistakes and can "understand" the vagaries commonly found in human thought.

A) Backward propagation
B) Forward propagation
C) Fuzzy logic
D) Neural computing

Question 22

Which of the following is not true of fuzzy logic?

A) Fuzzy logic focuses on gradation.
B) Fuzzy logic utilizes precise distinction.
C) Fuzzy logic allows for the partial description of a rule.
D) None of the above.

Question 23

Which of the following is an advantage of fuzzy systems?

A) Allows for increased association amongst the rules of the system
B) Supports modeling of contradiction
C) Increase system verification capability
D) All of the above.

Question 24

Which of the following is a disadvantage of fuzzy systems?

A) Fuzzy systems decrease system autonomy.
B) Objects in a fuzzy system can belong to multiple sets.
C) Fuzzy systems lack memory.
D) All of the above.

Question 25

A simple computer-based program whose primary function is to construct models of a problem space based upon trial and error is called a(n):

A) artificial neural network.
B) genetic algorithm.
C) decision support system.
D) data warehouse.

Question 26

In the human brain, the transmission of signals from one neuron to another occurs at a junction referred to as the:

A) transmitters.
B) neurides.
C) neurodes.
D) synapses.

Question 27

The ANN involves an interconnected system of nodes called _________that are associated with one or more weighted connections that are equivalent to human neural synapses, inside the memory of a digital computer.

A) transmitters
B) neurides
C) neurodes
D) synapses

Question 28

Which of the following is part of the basic structure of a neurode?

A) Bias input
B) State function
C) Nonlinear transfer function
D) All of the above.

Question 29

The ANN layer that receives the data is referred to as the?

A) input layer.
B) output layer.
C) hidden layer.
D) neural layer.

Question 30

The ANN layer that relays the final results of the net is referred to as the:

A) input layer.
B) output layer.
C) hidden layer.
D) neural layer.

Question 31

The __________ in an ANN is where the processing and transformation of the input signal takes place and where the type of output signal is determined.

A) input layer
B) output layer
C) hidden layer
D) neural layer

Question 32

The _____________ is not normally connected to the ANN and is assumed to have an input value of 1.0 for the state function. Its purpose is to allow for the individual adjustment of the firing threshold of the neurode to facilitate the final adjustment of the ANN following the learning process.

A) input layer
B) bias input
C) state function
D) output connection

Question 33

The __________________'s purpose is the consolidation of the weights of the various inputs to the neurode into a single value that can be passed to the transfer function for processing. The value obtained determines the degree of impact the combined inputs will have on the transfer function and, thus, on the final output of the neurode.

A) input layer
B) bias input
C) state function
D) output connection

Question 34

The function in an ANN which serves as the vehicle by which the summed information is passed on as output is called the:

A) transfer function.
B) state function.
C) memory function.
D) neural function.

Question 35

The actual procedure used by a neural network to find the appropriate weight settings is referred to as its:

A) transfer function.
B) state function.
C) weight seeking function.
D) learning paradigm.

Question 36

Which of the following is not a benefit derived from neural computing?

A) Reduced need for experts.
B) Allows for generalization from specific information content.
C) Highly verifiable, especially for complex problems.
D) All of the above are benefits.

Question 37

Which of the following are limitations of neural computing?

A) ANNs cannot explain inferences.
B) The repetitive training process is often time consuming.
C) Neural computing technologies push the limits of their hardware.
D) All of the above.

Question 38

A computer program that employs a set of adaptive processes that mimic the concept of "survival of the fittest" by regenerating recombinants of itself in response to a calculated difference between the network's guess and the desired solution state is known as a:

A) neural network.
B) artificial life system.
C) genetic algorithm.
D) None of the above.

Question 39

In a genetic algorithm, the smallest unit of information in the problem domain is called a:

A) chromosome.
B) gene.
C) neuron.
D) synapse.

Question 40

Which of the following is not a genetic algorithm operational process?

A) Evaluation
B) Initialization
C) Mutation
D) None of the above.

Question 41

Define neural computing.

Question 42

List the benefits of neural computing.

Question 43

List the limitations of neural computing.