Frågor baserade på tentan Flashcards

1
Q

Vad går all kunskap tillbaka på enligt David Hume?

A

Sannolikheter

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Sant eller falskt?
Aristoteles studerade resonemang som är logiskt korrekta i kraft av sin form

A

Sant

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Sant eller falskt?
John McCarthy myntade begreppet “machine learning”

A

Falskt. Han myntade begreppet artificial intelligence.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Hur många sensorer har Braitenbergfordon?

A

Två stycken

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Sant eller falskt?
En quiver plot använder färger för att visualisera funktionsvärden

A

Falskt

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Sant eller falskt?
En violin plot innehåller mer information än en box plot

A

Sant

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Kan tiodimensionella dataset representeras i kalkylark eller heat maps?

A

Kalkylark

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Vad är klassisk AI?

A

Klassisk AI kallas också ibland rule-based AI och bygger på förprogrammerade regler som agenten ska hålla sig till.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Vad är machine learning?

A

Machine learning är system som kan lära sig själva.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Bygger navigatorer på klassisk AI eller machine learning?

A

Klassisk AI

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Kan man studera vilka datapunkter som klassificerats rätt och fel i en confusion matrix?

A

Japp bror d kan man

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Vad är k-nn?

A

Står för k-nearest neighbours och är en maskinlärningsteknik som används vid klassificering och regression.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Sant eller falskt?
Vid stora värden på k tenderar k-nn att ge overfitting

A

Falskt!

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Vilket påstående är sant?
Välj ett alternativ:
Beslutsträd kan hantera numeriska features
Beslutsträd kan alltid konstrueras så att de ger 100% accuracy på träningsmängden

A

Beslutsträd kan hantera numeriska features

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

En mängd innehåller 2 lejon och 8 tigrar. Vad har den för Gini impurity?
Välj ett alternativ:
-0,52
-0.32
-0,42
-0.22

A

0.32

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Vilket påstående är sant?
Välj ett alternativ:
-Beslutsträd är ett exempel på en ensemble-metod
-Random forest bygger på majoritetsbeslut
-En random forest är enklare att tolka än ett beslutsträd

A

Random forest bygger på majoritetsbeslut

17
Q

Vilket påstående är sant?
Välj ett alternativ:
-Syftet med regularisering är att minska overfitting
-Regularisering är att man försöker minimera summan av de kvadrerade prediktionsfelen
-Outliers sorteras automatiskt bort av linjär regression

A

-Syftet med regularisering är att minska overfitting

18
Q

Vilken av dessa är en hyperparameter i DBSCAN?
Välj ett alternativ:
-Maximalt antal kluster
-Minsta antal punkter i ett kluster
-Minsta avstånd till grannar

A

-Minsta antal punkter i ett kluster

19
Q

Vilket av dessa påståenden om algoritmen k-means klustring är korrekt?
Välj ett alternativ:
-den använder sig av en slumpgenerator
-den returnerar en hierarki av kluster
-den bygger på att man flyttar datapunkterna närmare centralpunkterna
-den används för klustring i maximalt k dimensioner

A

den använder sig av en slumpgenerator

20
Q

Vad är hierarchial clustering?

A

The hierarchical clustering algorithm starts by considering each data point as a separate cluster and then iteratively merges the closest clusters together based on some distance metric. The algorithm continues merging the closest clusters until all data points belong to a single cluster or a predetermined number of clusters have been formed.

21
Q

Which of the following does both supervised learning and unsupervised learning require?
-input attribute
-output attribute
-hidden attribute

A

input attribute

22
Q

The output of training process in machine learning is _ _ _ _ _ _ _

A

machine learning model

23
Q

What is the maximum amount of classes if the input-output mapping is linearly separable?

A

2

24
Q

How many weights (excluding threshold/bias weights) does a fully-connected/dense ANN with two hidden layers consisting of 3 input nodes, 2 nodes in hidden layer 1, 2 nodes in hidden layer 2, and 1 output node, have?
Only one of the below is correct:
-8
-12
-18

A

12

25
Q

Given an input vector of x = [x1,x2] where x1 = 0.25, x2 = 1.0, and a weight vector w = [1.0, 0.5], what is the activation value as calculated by the dot product of x and w.
One correct answer:
-[0.25, 0.5]
-2.75
-0.75
-[1.25, 1.5]

A

0.75

26
Q

How many pooling maps does a standard ‘shallow’ Convolutional Neural Network (1 layer of feature maps + 1 layer of pooling maps) have that uses 4 filters of dimensionality 3x3 over the input image?
Choose one from the below
4
54
14

A

4 pooling maps. Anledning: det är 4 filters

27
Q

Which of the below correctly describe similarities and differences between convolution filters (used for encoding on feature maps) and pooling “filters” (used for encoding on pooling maps)?
Choose one from below
-Convolution filters are used for dot product computations and pooling filters are used for
max or mean computations
-Only convolutional filters ‘scan’ over their processed inputs
-Convolution filters are used for max or mean computations and pooling filters are used for
dot product computations

A

-Convolution filters are used for dot product computations and pooling filters are used for
max or mean computations

28
Q

Of the following, which is the best characterization of the Vanishing Gradient Problem
Choose one from below
-Backpropagating error/loss leads to increasingly small values for updating the weightsnearer to the input layer of the neural network.
-A problem for backpropagating error terms where the ANN is shallow

A

-Backpropagating error/loss leads to increasingly small values for updating the weights nearer to the input layer of the neural network.

29
Q

Which of the below best describes the term “local minimum” as compared to “global minimum”?
Choose one from below
-A point at which the network loss is not at its potential minimum
-A point at which the network loss is at its absolute minimum

A

-A point at which the network loss is not at its potential minimum

30
Q

Which of the following make CNNs efficient with respect to the number of parameters they haveas compared to “dense” ANNs?
Välj ett alternativ:
-Each feature map consists of a number of neurons whose encodings share a single filter of weights
-Adding convolutional layers to CNN always means having fewer parameters (weights) in
the full network

A

-Each feature map consists of a number of neurons whose encodings share a single filter of weights

31
Q

Which of the following are considered good stopping criteria when training ANNs?
Välj ett alternativ:
-Successive non decreasing loss/error for validation data checks
-Validation loss/error below validation loss

A

-Successive non decreasing loss/error for validation data checks

32
Q

How can you handle missing or corrupted data in a dataset?
Välj ett alternativ:
-Drop missing rows or columns
-Replace missing values with mean/median/mode
-All of the above

A

-All of the above

33
Q

Machine learning algorithms build a model based on sample data, known as
……………..

A

Training data

34
Q

Common classes of problems in machine learning is …………..
Välj ett alternativ:
Classification
Clustering
Regression
All of the above

A

All of the above

35
Q

True or false: Bootstrapping allows us to choose the same training instance several times.

A

True