Quiz
Neural Network
Create a free account to finish this quiz
Save your progress and see your final score.
Already have an account? Sign in
Questions in this quiz (133)
1 The primary aim of Machine Learning is to allow the computers learn automatically ...
- Awithout human intervention
- Bwithout data
- Cwithout computers
- Dall of them
2Supervised machine learning algorithms ...
- Acan apply what has been learned in the past to new data using labeled examples to predict future events
- Bare used when the information used to train is neither classified nor labeled
- Cfall somewhere in between supervised and unsupervised learning, since they use both labeled data for training - typically a small amount of labeled data and a large amount of unlabeled data
- Dis a learning method that interacts with its environment by producing actions and discovers errors or rewards
3What is Machine Learning?
- AIt is an application of artificial intelligence that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.
- Bit is a unique inter - neuron interface to transfer the information
- Cit is branched out into many, so-called terminals, which are terminated by a membrane to contact the thorns of dendrites of other neurons
- Dit is a nervous cell, which is the basic functional building element of nervous system
4Unsupervised machine learning algorithms ...
- Aare used when the information used to train is neither classified nor labeled.
- Bfall somewhere in between supervised and unsupervised learning, since they use both labeled and unlabeled data for training - typically a small amount of labeled data and a large amount of unlabeled data.
- Ccan apply what has been learned in the past to new data using labeled examples to predict future events
- Dis a learning method that interacts with its environment by producing actions and discovers errors or rewards.
5Machine Learning algorithms are often categorized as ...
- Asupervised or unsupervised
- Bpositive and negative
- Cwith computer and without computers
- Dall of them
6Reinforcement machine learning algorithms ...
- Ais a learning method that interacts with its environment by producing actions and discovers errors or rewards.
- Bfall somewhere in between supervised and unsupervised learning, since they use both labeled and unlabeled data for training - typically a small amount of labeled data and a large amount of unlabeled data.
- Ccan apply what has been learned in the past to new data using labeled examples to predict future events.
- Dare used when the information used to train is neither classified nor labeled.
7Machine Learning focuses on...
- Aconfiguration of the neural network
- Binter-neuron interface to transfer the information
- Cthe development of computer programs that can access data and use it learn for themselves
- Dneurophysiological principles to formulate a mathematical model of neural network
8Machine Learning ...
- Aenables analysis of massive quantities of data
- Ball of them
- Cdoesn't enable analysis of massive quantities of data
- Denables analysis of not massive quantities of data
9Combining machine learning with artificial intelligence and cognitive technologies can make it...
- Aeven more effective in processing large volumes of information
- Beven less effective in processing large volumes of information
10Semi-supervised machine learning algorithms ...
- Afall somewhere in between supervised and unsupervised learning, since they use both labeled and unlabeled data for training - typically a small amount of labeled data and a large amount of unlabeled data.
- Bare used when the information used to train is neither classified nor labeled.
- Ccan apply what has been learned in the past to new data using labeled examples to predict future events.
- Dis a learning method that interacts with its environment by producing actions and discovers errors or rewards.
11Output of the sigmoid/logistic activation function is always going to be in range
- Afrom 0 to 1
- Bfrom -1 to 1
- Cfrom 0,5 to 1
12Each connection (analogous to a synapse) between artificial neurons_______.
- Acan transmit a signal from one to another.
- Bis a neuron
- Ccan not transmit a signal from one to another.
- Dis a neural network
13Most popular type of activation functions - __________________
- Asigmoid or logistic
- Bstep-wise
- Chyperbolic tangent
- Dlinear
14.___________refers to the strength or amplitude of a connection between two nodes
- ASynaptic weight
- Bperceptron
- CActivation function
- DNeuron
15Each input__________by the weight associated with the synapse connecting the input to the current neuron.
- Ais multiplied
- Bis deleted
- Cis divided
- Dis classified
16 _________are connections between neurons.
- ASynapses
- BAxons
- CActivation functions
- DAll of them
17 A neuron takes a group of weighted inputs, applies__________, and returns an output.
- Aan activation function
- Ba nervous cell
- Ca biological neuron
- Da small amount of labeled data
18__________________is a learning algorithm that is inspired by the structure and functional aspects of biological neural networks
- AArtificial neural network learning algorithm
- BActivation functions
- CA nervous cell
- DBiological neuron
19In machine learning,________is an algorithm for supervised learning of binary classifiers (functions that can decide whether an input
not).- Athe perceptron
- Ban axon
- Ca neuron
- Dan activation function
20is a type of linear classifier
- APerceptron
- BAn axon
21Outputs in perceptron:
- A0 or 1
- BOnly 1
- CAny number
22 In machine learning, the perceptron is an algorithm for____________
- Asupervised learning
- Bunsupervised learning
- Csemi-supervised learning
- Dreinforcement learning
23The binary classifier defines that there should be only __ for classification:
- Atwo categories
- Bmany categories
- Cfive categories
- Done categorie
24The __ is a machine learning algorithm used to determine whether an input belongs to one class or another:
- Aperceptron
- Boutput
- CSigmoid function
- Dneuron
25Weights are initialized at __ values:
- Asmall random
- Blarge random
- Conly negative
26___ neuron allows a classifier to shift the decision boundary left or right:
- AA categorie
- BA weight
- CA bias
- DAn activation function
27Inputs in perceptron:
- AOnly 1
- BAny number from 0 to 1
- CAny number
- D0 or 1
28 In Learning Algorithm for perceptron we use:
- Astep function
- BSigmoid function
- Clinear function
29For linear classification it becomes necessary to make the perceptron as __ as possible.
- Alinear
- BAll of them
- Cnonlinear
30Backpropagation is_________
- Aa supervised learning
- Bsemi-supervised learning
31The training data set consists of input signals (xj and x2) assigned with______
- Acorresponding target (desired output)
- Bactivation function
- Cweights
- Da bias
32In each iteration_____________of nodes are modified using new data from training data set.
- Aweights coefficients
- Bactivation functions
- Cdesired outputs
33In backpropagation we use__________
- Asigmoid function
- Bstep function
- Clinear function
34Nodes that are no target of any connection are called
- Ainput neurons
- Boutput neurons
- Chidden neurons
35In one of algorithm step the output signal of the network у is compared with________
- Athe desired output value
- Bneuron
- Csynaptic weight
36The backpropagation algorithm has been worked out in__
- Athe 1980s
- Bthe 1960s
- Cthe 1940s
37If in backpropagation algorithm propagated errors came from few neurons they are_____
- Aadded
- Bmultiplied
38Most popular type of activation functions for backpropagation algorithm is____
- Asigmoid or logistic
- Blinear
- Cstep-wise
39To teach the neural network we need___________.
- Atraining data set
- Bmean pass phase
- Ca biological neuron
40____________computes 'functional signal', propagation of input pattern signals through network.
- AForward pass phase
- BBackward pass phase
- CMean pass phase
41The technique of backpropagation algorithm is used for_____
- Aall network layers
- Bonly the last layer
- Conly the first layer
42The training data set consists of______assigned with corresponding________
- Ainput signals, target
- Baxons,synapses
- Call of them
43The idea of backpropagation algorithm is to propagate________(computed in single teaching step) back to all neurons.
- Aerror signal
- Bsynaptic weight
- Cinput signal
44.___________computes 'error signal’, propagates the error through network starting at output units.
- ABackward pass phase
- BForward pass phase
- CMean pass phase
45Neural networks where the output from one layer is used as input to next layer are called ___ neural networks:
- Afeedforward
- Bstraightforward
- Cback forward
46The output layer of the network for recognition handwritten English letters contains ____.
- A26 neurons
- B1 neuron
- C33 neurons
47 Number of the neuron is this hidden layer of network _____:
- ACould be any number
- Bis in range from two to five
- Cis always equal to three
48There are no _______ in the feedforward network:
- Aloops
- Bnodes
- Cweights
49If the output of the network for recognition of handwritten digit was a 0 it means ____________:
- A"the input image is not a digit that this output indicate"
- B"the input image is a zero"
- C"error in the network"
50Artificial neural networks with feedback loops are called ___ neural networks:
- Arecurrent
- Bstraightforward
- Cfeedforward
51The output layer of the network for recognition handwritten digit with binary representation in the output contains _________:
- A4 neurons
- B10 neurons
- C1 neuron
52 The output layer of the network for recognition handwritten digit contains ______ :
- A10 neurons
- B1 neuron
- C9 neurons
53If training data for the network will consist of 28 by 28 pixel images and so the input layer contains _______neurons:
- A784
- B28
- C56
54May the output of the network for recognition of handwritten digit be less than 0,5. :
- AYes
- BNo
55Show different types of Optimization Algorithms used in Neural Networks? :
- AGradient Descent, Bath Gradient Descent, Stochastic gradient descent, Mini Batch Gradient Descent.
- Ball of them
- Cpositive and negative
- DWith computers and without computers
56Optimization Algorithm falls in 2 major categories:
- AFirst Order Optimization Algorithms and Second Order Optimization Algorithms
- Bpositive and negative
- Cwithout data and with data
- Dsupervised or unsupervised
57Gradient descent is a ______ iterative optimization algorithm :
- Afirst-order
- Bsecond-order
- Cthird-order
- Dnone of them
58What is a Cost Function?:
- AThe difference between the output value produced at the end of the Network and the actual value.
- BNeurophysiological principles to formulate a mathematical model of neural network
- CThe development of computer programs that can access data and use it learn for themselves
- DOptimization algorithm
59 When we train neural network we want to ____ cost function :
- Aminimize
- Bmaximize
- Cmove
- Dincrease
60What is Gradient descent? :
- AThe optimization algorithm used to minimize some function by iteratively moving in the direction of steepest descent as defined by the negative of the gradient.
- BThe learning method that interacts with its environment by producing actions and discovers errors or rewards
- CThe configuration of the neural network
- DThe difference between the output value produced at the end of the Network and the actual value.
61We use Gradient descent … :
- Afor finding the minimum of a function
- Bfor finding the maximum of a function
- Cfor processing large volumes of information
- Dall of them
62 MSE(for neural networks) is …:
- Amean of the squared differences between predicted y and actual y
- BMini Batch Gradient Descent
- CMicrosoft Security Essentials
- DMoneySavingExpert
63Cost function is…:
- Aloss or objective function.
- Bactivation function
- Cexponential function
- Dtrigonometric functions
64Are deep neural networks typically feedforward networks?
- AYes
- BNo
65The term Deep Learning was introduced to machine learning community in ____.
- A1986
- B1940
- C2003
66___ is a generative graphical model or alternative a class of deep neural network, composed of multiple layers of latent variables (“hidden units”), with connections between the layers but not between within each layer
- ADeep belief network
- BDeep neural network
- CArtificial neuron
- DSingle perceptron
67_____is an artificial network with multiple hidden layers between the input and output layers .
- ADeep neural network
- BBias
- CSingle perceptron
68Deep neural network, deep belief networks and neural networks are _____
- Adeep learning architectures
- Bartificial neurons
- Cperceptrons
69__________ is a part of broader family of machine based on learning data representations as opposed to task-specific algorithms
- ADeep learning
- BMachine learning
- CE-learning
70What is making deep networks hard to train?
- AThe different layers in deep networks are learning at vastly different speeds
- BModern algorithms
- CNothing
71Deep learning architectures have been applied to fields:___________
- AAll of them
- Bmachine translation
- Csocial network filtering
- Dspeech recognition
72The term Deep Learning was introduced to the machine learning community by __________
- ARina Dechter
- BGalina Pachshenko
- CDiana Kim
- DFrank Rosenblatt
73“To convolve” means …:
- Ato roll together
- Bto unroll
- Cto find out minimum
- Dto train
74Kernel in CNN is a… :
- Afilter
- Bconvolutional operator
- Cpool
- Dfunction
75We sometimes call the map from the input layer to the hidden layer …:
- Aa feature map
- Ba kernel map
- Ca hidden map
- Da weight map
76Pooling layers:
- Acombine the outputs of neuron clusters at one layer into a single neuron in the next layer
- Bcombine the outputs of neuron clusters at one layer into many neurons in the next layer
- Ccombine the outputs of neuron clusters at one layer into many pools
- Dcombine the outputs of neuron clusters at one layer into a pool on inputs
77CNNs use three ideas….:
- Alocal receptive fields, shared weights and polling
- Bconvolutional operator, activation function and cost function
- Cminimum, maximum and zero
- DNone of them
78Fully connected layers connect … :
- Aevery neuron in one layer to every neuron in another layer
- Bevery neuron in one layer to only one neuron in another layer.
- Cevery neuron in one layer to more than 50% neurons in another layer.
- Devery neuron in one layer to 50% neuron in another layer.
79Convolution is …:
- Aa mathematical operation on two functions to produce a third function that expresses how the shape of one is modified by the other.
- Ba neural network
- Ca bias
- Dgradient descent
80CNN (for neural networks) is …:
- AConvolutional Neural Network
- BCable News Network
- CClever Neural Network
- DControl Neural Network
81Each neuron in the first hidden layer in CNN will be connected to…:
- Asmall region of the input neurons.
- Bsmall region of the output neurons.
- Can output
- Danother CNN
82A convolutional network consists of … :
- Aan input and an output layer as well as multiple hidden layers
- Ban input and an output layer
- Conly one neuron
- D3 inputs, one neuron, one bias and 2 outputs
83Learning, in which the output neurons compete among themselves to be activated….
- Acompetitive learning
- Bactivated learning
- Csupervised learning
- Dunsupervised learning
84In SOM an activated neuron called a…
- Awinner-takes-all-neuron
- Bthe-winner-takes-it-all neuron
- Ctriggered neuron
- DABBA
85SOM(self-organizing map) was introduced by…
- ATeuvo Kohonen
- BFrank Rosenblatt
- CA.I.Galushkin
- DPamela McCorduck
86Lateral inhibition connections between the neurons are...
- Anegative feedback paths
- Bpositive feedback paths
- Cunchangeable feedback paths
- Dfeedback paths for biases
87The self organizing process involves four major components:
- AInitialization, Competition, Cooperation, Adaptation
- BProducing, Managing, Controlling,Finalization
- CSelf-organization,processing,involving,completing
- DBeginning, Initialization, Stopping, Ending
88The neuron whose weight vectors comes closest to the…
- Ainput vector
- Bself-organizing map
- Coutput vector
- Dmaximum
89Principal goal of an SOM is…
- Ato transform an incoming signal pattern of arbitrary dimension into a one or two dimension to perform this transformation adaptively in a topologically ordered fashion
- Bto find the cost function
- Cto find the winner
- Dto combine the outputs of neuron clusters at one layer into a single neurons in the next
90SOM (for neural network) is…
- Aa type of artificial neural network that is trained using unsupervised learning etc.
- BConvolutional Neural Network
- CNeural Network used in Turkic-speaking countries in Central Asia
- DSupervised Optical Map
91Kohonen Network is a…
- ASOM
- BCNN
- Cperceptron
- Dnetwork with Kohonen weights
92The winning neuron in SOM: the neuron whose weight vector lies closest to the input vector…
- AEuclidean distance
- Boutput of the neural network
- Cdifference
- Dtransforming an incoming signal pattern
93Connection in Hopfield network from a node to itself…
- Ais not allowed
- Bmust exist
- Cis a self-organizing map
- Dis allowed
94The Hopfield network ___________ "hidden" layer.
- Ahas no
- Bmust have
- Chas
- Dchanges
95in Hopfield networks have binary outputs (either 0,1 or -1,1)
- ANodes
- BCost functions in Hopfield networks
- CWeights
- DNetworks with Self-organization maps in Hopfield networks
96Hopfield network is…
- Arecurrent neural network (RNN)
- BConvolutional Neural Network
- CSOM(self-organizing map)
- Dperceptron
97The Hopfield network has…
- Asymmetrical weights with no self-connections
- Bsymmetrical weights with self-connections
- Casymmetrical weights with no self-connections
- Dasymmetrical weights with self-connections
98The an Hopfield network is RNN in which all connections are…
- Asymmetric
- Bnegative
- Ctriggered neuron
- Dasymmetric
99Hopfield networks were invented by…
- AJohn Hopfield
- BPamela McCorduck
- CJohn Hope
- DJohn Harrypotter
100Hopfield network is…
- Arecurrent neural network (RNN)
- Bperceptron
- CSOM(self-organizing map)
- DConvolutional Neural Network
101A recurrent neural network is....
- Aa class of artificial neural network where connections between units form a directed graph along a sequence.
- Ba class of self-organizing maps
- Ca Convolutional Neural Network
- Da class with winner takes all neurons
102Hopfield networks were invented in…
- A1982
- B1958
- C2003
- D1970
103Nodes in Hopfield networks are updated.
- Aasynchronously
- Bto find the cost function
- Csynchronously
- Dto find the winner
104The weight coefficients of the inhibitory synapses in the second layer in Hamming network are taken equal to
- A1 divided by number of inputs
- B1 divided by number of layers
- C1 divided by number of samples
- D1 divided by number of weights
105Inputs in Hamming network:
- A1 or -1
- B0 or 1
- Cany number from 0 to 1
- Dany number from -1 to 1
106The 1-st layer in Hamming network we use
- Aonly once
- Buntil output vector is stabilized
- Cuntil we find right weights
- Dfor many iteration
107The output of each neuron of is connected to input of each neuron of
- Athe output layer in Hamming network
- Bself-organization map
- Ca Kohonen network
- Dthe layer in Hopfield network
108Hamming network was introduced in…
- A1987
- B1915
- C1940
- D1982
109Connection in Hamming network from a node to itself...
- Ais allowed
- Bis not allowed
- Cis a self-organizing map
- Dmust not exist
110…. between two integers is the number of bits which are different at same position in both numbers.
- AHamming Distance
- BHopfield Distance
- CDistance from each weight to the sample vector
- DEuclidean distance
111The neuron synapse connected with its axon in Hamming network has a weight of…
- A1
- BT/2
- Cn/2
- D-1
112A Hamming network is...
- Aa two-layer network
- Ba one-layer network
- CConvolutional Neural Network
- Da multilayer network
113Hamming network was invented by
- ARichard P. Lippmann
- BRichard Hamming
- CJohn Hopfield
- DJohn Hamming
114The __________ is a neuron with N inputs and only one output
- Ainstar
- Boutstar
115____________a self organizing, competitive network based on unit
- AGrossberg Network
- BArtificial neuron
- CConvolutional Neural Network
116Synapses of neurons in the case of an input star perform the function of ____________ the signal
- Aclipping
- Badding
117The output of the _____________ is defined as weighted sum of its inputs
- Ainput star
- Boutput star
118One of the type of Recurrent Neural Networks is the ___________
- AGrossberg Network
- BPerceptron
- CArtificial neuron
119The _____________ is a neuron having one input and N outputs
- Aoutput star
- Binstar
120The __________ has an additional function: it generates the required excitation signal for other neurons whenever it is excited
- Aoutput star
- Binstar
121 ________________ is any network whose neurons send feedback signals to each other
- ARecurrent neural network
- BPerceptron
- CConvolutional Neural Network
122Grossberg network is a artificial neural network introduced in ____________?
- A1969
- B2017
- C1910
123Grossberk network is a artificial neural network introduced by ___________
- AStephen Grossberg
- BGalina Pachshenko
- CTeuvo Kalevi Kohonen
124The ____________ is a neuron with N inputs and only one output
- Ainstar
- Boutstar
125The role of the ______________ in counterpropagation network is to produce the pattern corresponding to the category output by the middle layer.
- Aoutput layer
- Binput layer
126By combining Kohonen learning and Grossberg learning a new type of mapping neural network is obtained
- AThe counterpropagation network
- BRecurrent neural network
- CConvolutional Neural Network
127A bidirectional Counterpropagation network is capable of a two-way
- Amapping
- Binput
128In counterpropagation network the hidden layer is a _______________with unsupervised learning.
- AKohonen network
- BArtificial neuron
- CConvolutional Neural Network
129In counterpropagation network the output layer is a ______________ fully connected to the hidden layer.
- AGrossberg (outstar) layer
- BConvolutional Neural Network
- CHopfield network
130Training in counterpropagation network is a ___________ procedure.
- Atwo-stage
- Bmulti-stage
131Counterpropagation network was proposed in _______________
- A1986
- B2018
- C1939
132___________________ can also be viewed as a bidirectional associative memory
- Acounterpropagation network
- Boutstar
133The _______________ in counterpropagation network uses a supervIsed learning procedure.
- Aoutput layer
- Binput layer
Information
- Description
- Quiz on Neural Network, created by Marzhan Dunbaeva on 10/01/2021.
- Author
- Marzhan Dunbaeva
- Language
- English
- Views
- 111
- Rating
- No ratings yet Rate this quiz