Week 1 : Assignment 1
Which answer is the most appropriate description of the concept of Recursion
the reduction of the solution to a problem into a sequence of solutions to sub-problems
the reduction of the solution to a problem into a parallell solutions to sub-problems
the decomposition of a problem into sub-problems followed by the solution of these sub-problems
the reduction of the solution to a problem into the the same form of solutions to sub-problems
Accepted Answers:
the reduction of the solution to a problem into the the same form of solutions to sub-problems
Which is the most appropriate description of the concept of Tensor ?
A matrix with vectors as components
A multidimensional table that is a n-dimensional generalization of scalars, vectors and matrices
A matrix with more and two dimensions
Accepted Answers:
A multidimensional table that is a n-dimensional generalization of scalars, vectors and matrices
Which is the most appropriate characterization of a Hamiltonian graph?
A graph that can be drawn in a plane without any edges crossed
A graph for which exists a closed path that traverses the vertices of the graph exactly once
A graph for which exists a closed path that traverses the edges of the graph exactly once
Accepted Answers:
A graph for which exists a closed path that traverses the vertices of the graph exactly once
Which is the most appropriate characterization of the concept of Convolution?
The recursive application of several functions on an argument
A mathematical operation which when applied to two functions (f and g) produces a third function.
An integral where the bounds are functions
Accepted Answers:
A mathematical operation which when applied to two functions (f and g) produces a third function.
Which is the most appropriate characterization of a Multivariate function
a function with properties defined by another function
a function that assumes two or more values for at least one argument value
a function of more than one variable
Accepted Answers:
a function of more than one variable
Which is the most appropriate characterization of the concept of Regression?
A technique for estimating relationships between variables
A technique for infering causes from effects
A technique for reducing compounds onto components
Accepted Answers:
A technique for estimating relationships between variables
Who pioneered work on Computer Vision in 1957?
Marvin Minsky
Ray Solomonoff
Oliver Selfridge
John Holland
Herbert Simon
Frank Rosenblatt
Accepted Answers:
Frank Rosenblatt
Who coined the term Machine Learning in 1959?
Marvin Minsky
John McCarthy
Allen Newell
Arthur Samuel
Herbert Simon
Frank Rosenblatt
Accepted Answers:
Arthur Samuel
To which category of search strategies does Depth First Search belong?
Brute-Force Search Strategies
Informed (Heuristic) Search Strategies
Local Search Algorithms
Accepted Answers:
Brute-Force Search Strategies
In knowledge representation, the collections of definitions of basic entities and their relationships for a specific domain is
referred to by a specific term. Which term?
Epistemology
Ontology
Phenomenology
Accepted Answers:
Ontology
The ´Logic Theorist´ developed in 1955 has been called ´The first Artificial Intelligence Program´. Which of the early AI
scientist was one of the co-developers.
Marvin Minsky
John McCarthy
John Holland
Allen Newell
Accepted Answers:
Allen Newell
On which calculus did John McCarthy base the LISP programming language?
FUNCTIONAL CALCULUS
DIFFERENTIAL CALCULUS
LAMBDA CALCULUS
Accepted Answers:
LAMBDA CALCULUS
What is MINIMAX?
A Decision Rule that maximizes the potential win in a best case scenario
A Decision Rule that minimizes the potential loss in a worst case scenario
A Decision Rule that optimizes the balance potential loss and potential win in worst and best cases
Accepted Answers:
A Decision Rule that minimizes the potential loss in a worst case scenario
What do you call a function that estimates the cost of a path from the current node in a search graph to the closest goal
node?
MINIMALISTIC
OPPORTUNISTIC
HEURISTIC
Accepted Answers:
HEURISTIC
Week 2 : Assignment 2
What does the term Feature Extraction refer to?
Deriving a feature from an analogue input like an image
Making a selection of a subset of features from a featureset
Creating a new set of features by transformation of a featureset
Accepted Answers:
Creating a new set of features by transformation of a featureset
´Which important class of machine learning algorithms X is defined as follows?
X is concerned with how agents take actions in an environment and changing state so as to maximize some notion of
cumulative reward.
Explanation-based learning
Inductive logic programming
Deep learning
Similarity-based learning
Reinforcement learning
Accepted Answers:
Reinforcement learning
What is meant by a NEAR MISS in Machine Learning
A misclassification of an example in Conceptual Clustering
A non optimal chosen hypotheses in supervised learning due to a non optimal selction and ordering of examples
An erroneous action in Reinforcement learning
A deliberately choosen negative examples close to positive examples
Accepted Answers:
A deliberately choosen negative examples close to positive examples
Choose the most adequate pair of characteristics of a feature that forms the basis for removing this feature?
Irrelevance and High Feature value Cardinality
High Feature value Cardinality and Redundancy
Irrelevance and Redundancy
High Feature value cardinality and Non-informativeness
Accepted Answers:
Irrelevance and Redundancy
Learning Bias is one of the core concepts in the classical frameworks for Inductive Learning. Which of the following
interpretations is the most adequate for capturing that concept?
Deliberately using a very simplified and restricted Hypothesis Language
Deliberately choosing particular subsets of examples as the basis for learning
Embed background knowledge in the definition of the hypothesis language.
Accepted Answers:
Embed background knowledge in the definition of the hypothesis language.
What is the term best suited for the situation depicted in the figure below, where the green line estimates the classification of
the instances?
Linear separability
Overfitting
Linear non-separability
Underfitting
Accepted Answers:
Overfitting
A taxonomy is a directed acyclic graph (DAG) under the Specialization relation. Which of the following Structures is NOT a
valid DAG
Accepted Answers:
A taxonomy is a category structure, typically a hierachical one. Normally features are attributed to all categories on the
different levels of the taxonomy. In this context, what is meant by INHERITANCE?
That a leaf (terminal) category is characterized by the aggregation of attributes of itself and all its ancestor categories.
That a category is characterized by the same attributes as its parent category only.
That a category is characterized by the same attributes as its sibling categories.
Accepted Answers:
That a leaf (terminal) category is characterized by the aggregation of attributes of itself and all its ancestor categories.
For everyday entities such as tools, clothes, furniture, fruits, pets etc, a typical taxonomy (category structure) has three
levels of categories : superordinate, basic and subordinate. What is the primary characteristic of the subordinate level.
It is the most abstract level with the fewest features
It is the level with the highest density of attributes and the most common sense and everyday character of features
It is the most concrete level with detailed features
Accepted Answers:
It is the most concrete level with detailed features
For everyday entities such as tools, clothes, furniture, fruits, pets etc, a typical taxonomy has three levels: superordinate,
basic and subordinate. Which of the following Categories in a fruit taxonomy would normally be characterized as
SUPERORDINATE?
Fruit
Apple
Macintosh
Accepted Answers:
Fruit
Which properties does a Penguin have according to the above taxonomy with a standard specialization relation?
Cannot Fly
Cannot Fly, Has a beak, Can control body temperature in relation to environment, Has a spinal cord, Can move
Cannot Fly, Has a beak, Can control body temperature in relation to environment, Has a spinal cord
Cannot Fly, Has a beak, Can control body temperature in relation to environment
Cannot Fly, Has a beak
Accepted Answers:
Cannot Fly, Has a beak, Can control body temperature in relation to environment, Has a spinal cord, Can move
Week 3 : Assignment 3
How are the 14 dataitems in the provided dataset distributed on the leafs of the decision tree above?
a=2 b=3 c =4 d=2 e=3
a=3 b=2 c =4 d=3 e=2
a=3 b=2 c =3 d=2 e=2
Accepted Answers:
a=3 b=2 c =4 d=3 e=2
What is the normal notation for the probability of the occurence of one event A given another event B ?
P(B) AND P(A)
P(B,A)
P(A)<-P(B)
P(A|B)
Accepted Answers:
P(A|B)
Which is the correct description of the following Bayesian Network in terms of basic subnetworks?
Sequence, Convergence, Divergence
Sequence, Divergence, Convergence
Divergence, Sequence, Convergence
Divergence, Convergence, Sequence
Convergence, Sequence, Divergence
Convergence, Divergence, Sequence
Accepted Answers:
Convergence, Sequence, Divergence
A version of bayes rule: P(cause|effect) = P(effect|cause)* P(cause)/P(effect)
In this example
effect = person has a fever
cause = person has a flu
Prior probabilities: P(cause) = 0.0001, P(effect)= 0.002
P(effect|cause) = 0.8
What is the value of P(cause|effect)
0.08
0.016
0.04
Accepted Answers:
0.04
Which of the following pairs of steps is not consistent with the normal sequence of steps in Genetic Algorithms?
Generation, Selection,
Selection, Evaluation
Evaluation, Mutation
Evaluation, Reproduction
Reproduction, Mutation
Accepted Answers:
Evaluation, Mutation
Two chromosomes
A
B
01110000
10001111
A two point crossover operation in-between positions 1 and 8 should be performed
to give two new chromosomes. Which constellation is the crossover result?
10001111
01110000
00001110
11110001
11110001
00001110
Accepted Answers:
00001110
11110001
What is meant by The Bucket Brigade Phase of a Classifier system based on Genetic Algorithms?
The Bucket Brigade phase involves an algorithm for handover of results forwards along a sequence of rule
applications.
The Bucket Brigade phase involves an algorithm for handover of results forwards between two steps in a sequence of
rule applications.
The Bucket Brigade phase involves an algorithm for attributing credit and blame backwards along a sequence of rule
applications.
The Bucket Brigade phase involves an algorithm for attributing credit and blame backwards between two steps in a
sequence of rule applications.
Accepted Answers:
The Bucket Brigade phase involves an algorithm for attributing credit and blame backwards along a sequence of rule
applications.
What is the term for the points of connectivity between neurons?
AXON
DENDRITE
SOMA
SYNAPSE
Accepted Answers:
SYNAPSE
Which is the best estimate of the number of neurons in the human brain?
1 Billion
10 Billions
100 Billions
1000 Billions
Accepted Answers:
100 Billions
What is the output value Y for the above artificial neuron?
0.9
0.97
0.72
Accepted Answers:
0.72
Week 4 : Assignment 4
The different data-driven search approaches to induction require different storage policies.
What does the Breadth First Search algorithm have to store?
All positive examples + All negative examples + All hypotheses
All positive examples + all hypotheses
All negative examples + all hypotheses
All hypotheses
Accepted Answers:
All negative examples + all hypotheses
An inductive system learns from unordered pairs of dataitems of the form (size, color, animal)
where size= (large or small), color=(black or brown) and animal=(cat or dog).
How does the set of hypotheses look like after having handled the following three instances:
Common Instructions for Q2-Q4
Instance 1. { (Large Black Dog) (Small Brown Cat) }
Instance 2. { (Large Brown Cat) (Small Black Dog)}
Instance 3. { (Large Brown Dog) (Small Brown Dog) }
positive
positive
negative
Using Depth First Search
CBH3 ={ (Large ? ?) (Small ? Cat) }
CBH3={ (Large ? Dog) (Small ? ? ) }
CBH3=(? Black Dog) (? Brown Cat)}
Accepted Answers:
CBH3=(? Black Dog) (? Brown Cat)}
Using Breadth First Search
S3 = { (? Black Dog ) (? Brown Cat) }
S3 ={ (Large ? ?) (Small ? Cat) }
S3={ (Large ? ?) (Small ? ? ) }
Accepted Answers:
S3 = { (? Black Dog ) (? Brown Cat) }
Using the Version space approach
S3 ={(Large ? ?) (Small ? Cat) } G3 = { (? Black ?) (? ? Cat) } { (? ? Cat) (? Brown ?) }
S3 ={ (? Black Dog) (? Brown Cat) } G3 = { (? Black ?) (? ? ?) } { (? ? Cat) (? ? ?) }
S3 ={S3={ (Large ? ?) (Small ? ? ) } G3 = { (? Black ?) (? Brown ?) } { (? ? Cat) (? ? Dog) }
Accepted Answers:
S3 ={ (? Black Dog) (? Brown Cat) } G3 = { (? Black ?) (? ? ?) } { (? ? Cat) (? ? ?) }
How does the set of hypotheses look like after having handled the following three instances:
Common Instructions for Q5-Q7
Instance 1. { (Large Black Dog) (Small Brown Cat) }
Instance 2. { (Large Brown Cat) (Small Black Dog)}
Instance 3. { (Large Brown Dog) (Small Brown Dog) }
Using Depth First Search
CBH3 ={ (Large ? ?) (Small ? Cat) }
CBH3={ (Large ? Dog) (Small ? ? ) }
CBH3=(? Black Dog) (? Brown Cat)}
Accepted Answers:
CBH3=(? Black Dog) (? Brown Cat)}
positive
positive
negative
Using Breadth First Se
S3 ={ (Large ? ?) (Small ? Cat) }
S3 = { (? Black Dog ) (? Brown Cat) }
S3={ (Large ? ?) (Small ? ? ) }
Accepted Answers:
S3 = { (? Black Dog ) (? Brown Cat) }
Using the Version space approach
S3 ={(Large ? ?) (Small ? Cat) } G3 = { (? Black ?) (? ? Cat) } { (? ? Cat) (? Brown ?) }
S3 ={ (Large ? ?) (Small ? ? ) } G3 = { (? Black ?) (? Brown ?) } { (? ? Cat) (? ? Dog) }
S3 ={ (? Black Dog) (? Brown Cat) } G3 = { (? Black ?) (? ? ?) } { (? ? Cat) (? ? ?) }
Accepted Answers:
S3 ={ (? Black Dog) (? Brown Cat) } G3 = { (? Black ?) (? ? ?) } { (? ? Cat) (? ? ?) }
Which of the following statements is TRUE for a decision tree?
An attribute with lower information gain should be preferred to other attributes.
The entropy of a node typically decreases as we go down a decision tree.
A Decision tree is an example of a linear classifier.
The Entropy of a set increases with its purity.
Accepted Answers:
The entropy of a node typically decreases as we go down a decision tree.
What is the entropy for a decision tree data-set with 6 positive and 4 negative examples.
0.58
0.840
0.41
0.97
Accepted Answers:
0.97
What is the value of the Information Gain in the following partitioning?:
0.26
0.38
0.42
0.18
Accepted Answers:
0.38
For which values of k is the query instance classified as RED by the k-nearest neighbour algorithm?
3 and 7
1 and 3
3 and 5
1 and 7
Accepted Answers:
3 and 7
For the two feature vectors (1,0,1,0,0,0,1) and (0,0,1,1,1, 1,1).
What is the Manhattan distance between the two vectors?
5
3
2
4
Accepted Answers:
4
Consider the following two vectors:
X = [4, 0, 0, 5, 1, 0, 0]
Y = [5, 5, 4, 0, 0, 0, 0]
What is the cosine similarity between X and Y?:
0.53
0.43
0.17
0.38
Accepted Answers:
0.38
Consider a two-dimensional non linearly separable input space and the following feature mapping onto a linearly separable
three dimensional space:
X: = F : ( x1, x2 ) -> (x1^2, x2^2, sqrt 2*x1*x2)
Z: = F : ( z1, z2 ) -> (z1^2, z2^2, sqrt 2*z1*z2)
K (X,Z)= the Kernel or Similarity function in the three dimensional space = the inner product of X and Z in that space : |X,Z|
Calculate K(X,Z) = the inner product of X and Z above for the two two-dimensional vectors of (3,2) and (2,3) in the input
space
6
144
124
12
Accepted Answers:
144
Partitioning–based clustering is one of the approaches to Cluster Analysis. Which of the following specific techniques falls in
this category?
CLIQUE
COBWEB
K-means
Chameleon
DBSCAN
Accepted Answers:
K-means
In hierarchical-based clustering, it is typical to work with Distance Matrices between data items and so called Proximity
Matrices between clusters. A distance Matrix for 6 data items is given below. We assume three clusters: A-B, C-D and E-F.
How does the corresponding Proximity Matrix between these clusters look like?
Accepted Answers:
Week 5 : Assignment 5
Explanation Based Learning typically takes a predefined Domain Theory and modifies it in such a way that new more
complex rules are created as specific combinations of several existing rules. The goal is to make problemsolving more
efficient and the rule formation process is guided by the problem examples considered for training. As part of the
generalization process for the newly constructed rules a specific kind of pattern matching technique is used. Which is the
established term for this technique?
Variable matching
Resolution
Recursion
Unification
Accepted Answers:
Unification
Domain Theory:
Safe to travel(P,A,C) :- Healthy(P), Error free(A), No epidemics (C).
Error free(A):- Passed an inspection (A). Errorfree(A):- Worked when driven same day (A). Healthy(P) :- Had no
symptoms for three weeks (P).
Healthy(P) :- Passed a health test(P) Passed a health test (X):- On line test (X).
Passed a health test (X):- Clinic test (X). Passed an inspection (X):- Company inspection(X).
Passed an inspection(X):- Authority inspection(X). No epidemics (C):-No WHO alert (C).
Operational predicates:
Worked when driven last same day, Had no symptoms for three weeks, On line test, Clinic test, Company
inspection, Authority inspection, No WHO alert
Training Example:
Safe to travel (John, IndiaAir2020, India) Passed a health test(John), Authority inspection(IndiaAir 2020), No WHO
alert (India).
Goal concept:
Safe to travel(P, A, C)
Explain and generalize from Safe To Travel( John, IndiaAir2020, India). How does the resulting rule look like after applying
an EBL algorithm
Safe to travel(John,IndiaAir2020,India)) :- Healthy(John), Error free(IndiaAir2020), No epidemics (India).
Safe to travel(John, IndiaAir2020, India):-Passed a health test (John), Authority inspection(IndiaAir 2020), No WHO
alert (India).
Safe to travel(P, A, C):-Passed a health test (P), Authority inspection(A), No WHO alert (C).
Safe to travel(P,A,C)) :- Healthy(P), Error free(A), No epidemics (C).
Accepted Answers:
Safe to travel(P, A, C):-Passed a health test (P), Authority inspection(A), No WHO alert (C).
Which of the following inference schemes correspond to Induction?
A
B
C
Accepted Answers:
B
ILP algorithm constructs a lattice of clauses fabricated from the building blocks (predicates, variables, constants) specific to
each induction case and navigates that lattice in order to establish a Hypothesis set that fits the current set of examples. ILP
can work in two fashions: Top Down and Bottom Up, establishing specialization or generalization relations in the lattice.
Case
Hypotheses: insect(X).
Positive Examples:
insect(silverfish), insect(green drake), insect(cricket)
Background knowledge:
Has six legs (fire brat), Has six legs (green drake), Have wings(green drake), Has six legs (cricket), Have
wings(cricket), Can fold wings(cricket),
How does the Hypotheses set of clauses look like after a first iteration of a Top Down ILP algorithm based on the first
positive example insect(fire brat) ?
insect(X):-Has six legs(X).
insect(X):-Has six legs(X), insect(X):-Has wings(X), insect (X):-Can fold wings(X).
insect(fire brat):-Has six legs(fire brat).
Accepted Answers:
insect(X):-Has six legs(X), insect(X):-Has wings(X), insect (X):-Can fold wings(X).
Instance Based Learning techniques are typically used in one of the main phases of Case Based Reasoning. Which?
RETRIEVE
REUSE
RETAIN
REVISE
Accepted Answers:
RETRIEVE
Choose the correct sequence of phases:
X Y Z W = RETRIEVE REUSE RETAIN REVISE
X Y Z W = RETRIEVE REVISE REUSE RETAIN
X Y Z W = RETAIN REVISE REUSE RETRIEVE
X Y Z W = RETAIN RETRIEVE REVISE REUSE
Accepted Answers:
X Y Z W = RETAIN RETRIEVE REVISE REUSE
A dynamic programming approach to calculating the shortest distance from S to T gives the result= 14.
Which result would a greedy forward search give?
26
21
14
Accepted Answers:
26
Monte Carlo Reinforcement Learning MC methods learn directly from samples of complete episodes of experience.
•
•
First-visit MC: average returns only for first time s is visited in an episode.
Every-Visit MC: average returns for every time s is visited in an episode.
Simplified Algorithm for first-visit Monte Carlo
1. Initialize state-value functions. Return list(s) ← empty list
2. For all s in all episodes E
3. return = sum of rewards r in episode from state to termination.
4. if this is the first occurrence of this state s add the calculated return to the returns list (s).
5. Calculate values of all s as average over all return lists (s).
Example: An undiscounted Markov Reward Process with two states A and B. The Transition matrix and reward function are
unknown. Two sample episodes: E1 A->A, r=2 A->B, r=7 B->A, r= -7 A->B, r=7 B->terminate, r= -2. E2 B->A, r= -7 A->B, r=7
B->terminate, r = -2
Which are the estimated values of A and B after considering the two episodes using the simplified first-visit algorithm?
V(A)= 6,V(B)= -2
V(A)= 7,V(B)= -2
V(A)= 6,V(B)= 2
V(A)= 7,V(B)= 2
Accepted Answers:
V(A)= 6,V(B)= -2
In this example, we have a 4X5 board. The start state is marked above with a blue dot. The rewards for all positions are also
marked (rewards being 1, -4 and 4). The positions on the rim of the board are terminal states. The board elements are
labelled from 1 to 20 as in the mid figure. Moves are N, S, E and W. We consider two episodes: 1. 12:E->13:E->14:E->15
and 2. 12:N->7:E->8:E->9:E->10. What are the updated versions of the Q elements for state 12, 13, 14, 7,8 and 9 after
considering these two episodes? α=γ=1α=γ=1.
12,E=0 13,E=0, 14,E= -4
12,N=0, 7,E=0 8,E=0 9,E = 4
12,E=1 13,E=1 14,E= -6
12,N=1 7,E=1 8,E=1 9,E= 6
12,E=1 13,E=2 14,E= -6
12,N=1 7,E=2 8,E=3 9,E= 6
12,E=1 13,E=2 14,E= -2
12,N=1, 7,E=2 8,E=3 9,E=7
Accepted Answers:
12,E=1 13,E=1 14,E= -6 12,N=1 7,E=1 8,E=1 9,E= 6
Week 6 : Assignment 6
One of the logical operators described by the following truthtables is infamous for being not linearly separable and therefore
not possible to implement by a single perceptron.
Which?
Accepted Answers:
How will the final weight vector look like when all data-items are processed
1011
0 0 -1 0
0 -1 0 0
1010
None of the above
Accepted Answers:
0 -1 0 0
In a feedforward ANN, the so called Delta Rule is based on a particular error measure based on the target values and output
values for the outputs of the ANN.
When this particular error measure is applied for a single neuron, what is the error measure for a target value of 12 and an
output value of 14.
E= ½ * ( T-Y) ^2
4
2
0.5
16
Accepted Answers:
2
How will the final weight vector look like when all data-items are processed?
-0.25 0 0 0.25
0 -0.25 0.25 0
0.5 0 0.75 -0.35
-0.25 0.5 -0.25 0.25
None of the above
Accepted Answers:
-0.25 0.5 -0.25 0.25
What is the output value Y from neuron # 6 ?
0.118
0.018
1.118
1.018
Accepted Answers:
0.018
What is the backpropagated error sensitivity at neuron # 1 ?
0.0144
0.040
0.004
0.014
Accepted Answers:
0.004
What is the updated weight for connection between input x1 and neuron #1?
0.104
0.140
0.1144
0.114
Accepted Answers:
0.104
Select the main reason for extending standard RNNs to so called Bidiectional RNNs:
being able to handle multiple levels of abstractions within the domains of analysis
being able to get information from past (backwards) and future (forward) states simultaneously
Mastering the vanishing gardient problem
Mastering the long dependency problem
Accepted Answers:
being able to get information from past (backwards) and future (forward) states simultaneously
Which RNN network structure would best fit a text analysis task, where the occurences of references to a specific kind of
event are searched for?
Accepted Answers:
Hebb’s Law can be represented in the form of two rules:
1.
2.
If two neurons on either side of a connection (synapse) are activated synchronously, then the weight of that
connection is increased.
If two neurons on either side of a connection (synapse) are activated asynchronously, then the weight of that
connection is decreased.
Learning according to Hebb’s Law is primarily consistent with one of the following kinds of learning
Reinforcement learning
Un-supervised learning.
Supervised learning
Accepted Answers:
Un-supervised learning.
How will the final weight vector look like when the training data has been processed?
3223
2233
3322
3232
None of the above
Accepted Answers:
3322
Hopfield Networks use a method borrowed from Mettalurgy to let the network states settle. The method can be characterized
as of below:
•
•
•
Heat the solid state metal to a high temperature
Cool it down very slowly according to a specific schedule.
If the heating temperature is sufficiently high to ensure random state and the cooling process is slow enough
to ensure thermal equilibrium, then the atoms will place themselves in a pattern that corresponds to the
global energy minimum of a perfect crystal.
What is the name of the method
Casting
Crystallization
Annealing
Welding
Forging
Accepted Answers:
Annealing
What is the resulting state vector ?
1 -1 1
1 1 -1
-1 1 1
111
Accepted Answers:
111
how does the feature map look like after the convolution of input array A with the filter F?
Accepted Answers:
Week 7 : Assignment 7
A movement within Psychology has also influenced re-inforcement Learning. Which?
Behaviourism
Cognitivism
Reductionism
Structuralism
Functionalism
Accepted Answers:
Behaviourism
A movement within Psychology is consistent with an abstract model of the Mind in terms of systems of Symbols, Concepts
and Logically related inferences. Which movement?
Reductionism
Behaviourism
Structuralism
Cognitivism
Functionalism
Accepted Answers:
Cognitivism
A movement within Psychology is consistent with the motivations behind the sub-symbolic approach of artificial neural
networks. Which movement?
Cognitivism
Behaviourism
Structuralism
Reductionism
Functionalism
Accepted Answers:
Reductionism
Learning of Language and language engineering in CS and AI is strongly inspired by the structuralist view on Linguistics.
Who was the front figure for this view?
Bloomfield
Hockett
Lakoff
Chomsky
Accepted Answers:
Chomsky
What is the neuro science term for the property of a system to adapt, reconfigure and consolidate its structure?
Locality
Plasticity
Equipotentiality
Mass action
Specificity
Holism
Accepted Answers:
Plasticity
The work by Donald Hebb exemplifies an integration of two of the principles below. Which two?
Locality, Specificity
Equipotentiality, Mass Action
Locality, Holism
Mass action, Plasticity
Resilience, Plasticity
Resilience, Equipotentiality
Accepted Answers:
Locality, Holism
The Classical view of categories is characterized by the following statements. One statement is wrong. Which?
Categories are arbitrary. Typically the ways we choose to categorize objects are culturally and linguistically based
The levels in a hierarchy (lattice) of categories have a different status , with the middle layer being most fundamental to
everyday thinking (basic level) in contrast to more abstract (superordinate) and more detailed (subordinate) levels
Categories have defining attributes (features/value combinations). All members share them, No nonmembers share
them and there are no overlap between members and nonmembers
The Intension (set of attributes) determines the extension of a category (its members)
The member space has no internal structure and all members are regarded as equal and first class citizens.
Accepted Answers:
The levels in a hierarchy (lattice) of categories have a different status , with the middle layer being most fundamental to
everyday thinking (basic level) in contrast to more abstract (superordinate) and more detailed (subordinate) levels
The Modern or natural view of a categories is characterized by the following statements. One statement is wrong. Which?
Categories are in many case motivated by properties of our sensory system and the world surrounding us.
Categories are built around central members or prototypes defined by sharing more attributes with other members than
with nonmembers. Members are graded based on typicality
Obviously typicality generates a topology of the member space
The Intension (set of attributes) crisply determines the extension of a category (its members)
The levels in a hierachy (lattice) of categories have a different status , with the middle layer being most fundamental to
everyday thinking (basic level) in contrast to more abstract (superordinate) and more detailed (subordinate) levels
Accepted Answers:
The Intension (set of attributes) crisply determines the extension of a category (its members)
The Entropy measure is used to guide the optimal setup of a decision tree. The measure is inspired by entropy in physics
which measures the:
Molecular regularity
Molecular disorder
Molecular non-regularity
Molecular order
Accepted Answers:
Molecular disorder
The Energy function in Hopfield Networks is inspired by a model of ferromagnetism in Statistical Mechanics. The model is
named after a person. Which?
Lenz
Hopfield
Ising
Onsager
Accepted Answers:
Ising
Which Dataset Repository is primarily a gate-way to other specific repositories for Data-sets
Imagenet
Kaggle Repository
Google trends
Amazon reviews
Accepted Answers:
Kaggle Repository
Deep Mind is a company focussing on Artificial Intelligence being related to another large company. Which?
Amazon
Microsoft
SAP
Google
Facebook
Huawei
Accepted Answers:
Google
What does the abbreviation API stands for?
Advanced Programming Interface
Application Programming Interface
Advanced Project Integrator
Accepted Answers:
Application Programming Interface
Which of the following tools are NOT Open Source?
Watson
TensorFlow
Python
Wolfram
Caffe
Accepted Answers:
Wolfram
Which of these systems for technical computations is primarily aimed for numerical calculations in contrast to symbolical?
Mathematica
Macsyma
MATLAB
Maple
Reduce
Accepted Answers:
MATLAB
Which of the programming languages below can be characterized as a ´Multi Paradigm Language´ in the sense that it
supports both imperative, functional, object oriented programming and indirectly logic oriented programming?
Pascal
C
C++
C#
LISP
Haskel
Prolog
Python
Wolfram
R
Java
Java Script
Visual Basic
Accepted Answers:
Python
One popular tool in a currently typical machine learning toolbox, provides collaborative environments, where you can freely
combine human-readable narratives with computer-readable codes, equations and visualizations. Pick this tool from the list
below:
Caffe
Pytorch
Theano
Jupyter Notebooks
Tensorflow
Microsoft Cognitive Toolkit
Accepted Answers:
Jupyter Notebooks
Many Machine Learning tools are specialized for support of ANN, in particular Deep Learning. Examples are given below.
One of the listed tools have an ambition for support for Machine Learning techniques in general. Which one?
Intel deep learning cloud
Microsoft Cognitive Toolkit
Tensorflow
Caffe
Accepted Answers:
Tensorflow
Which two of the following platforms are primarily NOT aimed for Cloud/Distributed Computing?
Amazon Webservices
Apple iCloud
SAP Leonardo
Microsoft Azur
Hadoop MapReduce
Apache SPARK
Accepted Answers:
SAP Leonardo
The dominating hardware providers are launching Systems on a Chip (SoCs) that can provide the basic functionalities for
Advanced Driver Assist Systems (ADAS) on various levels (2-5). The core of the functionality of the SoCs is the image
recognition enabled by Machine Learning (Deep Learning)
Which is the correct coupling between the hardware providers and ADAS/SoCs?
Qualcomm - Drive Xavier/Pegasus
Nvidia - Snapdragon Ride
Qualcomm - Drive Xavier/Pegasus
Nvidia - Mobileeye
Qualcomm - Snapdragon Ride
Nvidia - Drive Xavier/Pegasus
Intel - Mobileye
Intel - Snapdragon
Ride
Qualcomm - Snapdragon Ride
Xavier/Pegasus
Nvidia - Mobileye
Qualcomm - Mobileye
Xavier/Pegasus
Nvidia - Snapdragon Ride
Qualcomm - Mobileye
Nvidia - Drive Xavier/Pegasus
Intel - Mobileye
Intel - Drive
Intel - Drive
Intel - Snapdragon Ride
Accepted Answers:
Qualcomm - Snapdragon Ride
Nvidia - Drive Xavier/Pegasus
Intel - Mobileye