Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

FITE1010 > AI & ML Lect6 > Flashcards

AI & ML Lect6 Flashcards

1
Q

What is general AI?

A

Complex machines that possess same characteristics of human intelligence

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

What is Narrow AI?

A

Tech that are able to perform specific tasks, better than humans can

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

What is a turing test?

A

A test of a machine’s ability to exhibit intelligent behaviour indistinguishable from that of a human.

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

CAPTCHA is an example of?

A

Reverse Turing Test

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

AI in 1960’s used?

A

If then else which was pre defined rule and knowledge

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

Modern AI used ? to replace defined rule and knowledge

A

Machine learning algo

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

Use of Algo

A

Parse data, learn and make a determination or prediction

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

What is ML

A

Getting computers to act without being explicitly programmed

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

What is Deep Learning

A

Artificial neural networks have discrete layers, connections, and directions of data propagation

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

Steps in Data Science

A

Data collection, Data preparation, EDA (explanatory data analysis), Machine learning, Visualization

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

Why the need for Machine Learning?

A

No human experts, blackbox human expertise, rapidly changing phenomena, need for customization/ personalization

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

Give an example of Learning

A

Data: loan application data
Task: predict whether loan should be approved or not
Performance measure: accuracy
No learning: approve all future transactions
With learning: analyze user’s assets and credit score etc

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

How to achieve good accuracy om test data?

A

training example must be similar to test data

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

Types of Machine Learning

A

Supervised Learning, Unsupervised Learning, Reinforcement Learning

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

Uses of Supervised Learning

A

Classification and Regression

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

Uses of Unsupervised Learning

A

Clustering, Dimensionality Reduction, Anomaly detection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Characteristics of Supervised learning

A

Train using labeled data, direct feedback, predict outcome

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Characteristics of Unsupervised learning

A

No labels, no feedback, no hidden structure, learn with their own

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

Characteristics of reinforcement learning

A

Decision process, reward system, learn series of actions, has a mapping structure that guides machine from input to output

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Goal of unsupervised learning

A

Find Pattern and trend to discover output, doesn’t predict/find anything specific

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is classifcation in supervised learning

A

Categorize all variable that form the output

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

Example of Classification

A

Classifiying written digits in a cheque

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

Use of Regression

A

Identifying/Predicting a specific value, usually a real number, used in stock market prediction, sales volume by using mathemathical functions

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

Use of Clustering

A

identifying groupings occuring within the data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

Use of Anomaly detection

A

identifying anomalies within the data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

Use of Dimensionlity reduction

A

Transformation of data form high to low by discarding redundant data, while still retaining meaningful properties of the original data

Uses PCA

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

How Supervised Learning works:

A
  1. Provide Algo with labeled input and output data to learn
  2. Feed the machine new unlabeled info to see if it tags new data approporiately.
  3. If not, continue refining the program
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

Use of classification

A

Sorting items into categories

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

How unsupervised learning works:

A
  1. Input raw data
  2. Interpretation
  3. Algo
  4. Processing
  5. Output
30
Q

How reinforcement learning works:

A
  1. input raw data
  2. enviroment (state of data, agent takes control, selection of algo, best action, reward if correct)
  3. output
31
Q

2 features in Dimensionality reduction

A

Feature selection and Feature extraction

32
Q

What is feature selection in Dimensionality reduction?

A

identify redundant features and discard

33
Q

What is feature extraction?

A

find new set of low dimensional point that represent the original data well

34
Q

What is PCA

A

Principal Component Analysis is a method to construct low-dimensional representation of the data by focusing on the principal components

35
Q

Condition of Data in PCA

A

As much variance as possible

36
Q

Use of PCA

A

Data visualization with a scatter plot

37
Q

Best classifier for text classification

A

SVM

38
Q

Which method performs best in classifying high dimensional data?

A

SVM

39
Q

What is Decision Tree mainly used for?

A

Classification due to its competitive accuracy, and very effecient

40
Q

Leaf nodes in Decision tress indicate?

A

Class / decision

41
Q

What type of Algo for decision Tree?

A

Heuristic Algo

42
Q

How to make an easy to understand and better performing decision tree

A

Smaller and accurate tree

43
Q

Finding the best tree is?

A

NP-hard

44
Q

What is good attribute selection in Decision Tree?

A

Attribute that splits examples into subsets that are ideallly all (+) or (-)

45
Q

What is ensemble learning

A

A set of decision tress working together to make a single predicition, and allow greater predictive accuracy

46
Q

What are “Random forests”?

A

A set of decision trees, with each tree having different features

47
Q

Key Idea of KNN (Key Neighbour Model)

A

Properties of an input x are likely to be similar to those points in the neighbourhood of x

48
Q

Basic idea of kNN

A

Find k nearest neighbors of x and find target attribute of x based on corresponding attribute values

49
Q

How does complexity grow with data

A

Linearly

50
Q

How does kNN define nearest neighbors?

A

Euclidian Distance

51
Q

Training algo in kNN

A

Add each training example (x,y) to dataset D

52
Q

Classification algo in kNN

A

Count the K-nearest neighbors

53
Q

kNN is slow at?

A

Classification time eventho accuracy can be quite strong

54
Q

Usage of kNN

A

handwritten character classifications, recommender systems, medical data mining, pattern recognition

55
Q

How Linear Regression works?

A

Have a set of points, which the regression algo will model relationship between a single feature (explanatory variable x), and a continuous valued response (target variable y)

56
Q

Goal of Linear Regression

A

Find a best fit line such that the cost function is minimized

57
Q

Most common cost function is

A

MSE (Mean squared error), which is average squared diff between an observation’s actual and predicted values

58
Q

1 method of clustering is?

A

k-means algo

59
Q

What is a training set?

A

A set of examples used for learning a model

60
Q

What’s a validation set?

A

Set of examples that can’t be used for learning the model but can help tune model parameters. Helps control overfitting.

61
Q

What’s a test set?

A

Used to access the performance of the final model and provide an estimation of the test error.

62
Q

What set to never use in tuning parameters or revise model?

A

Test set

63
Q

How to do cross validation

A

Train the model on p% of the data
Test the model on the other (100-p)% of the data - this data is unseen by the model

64
Q

ML in Fintech

A

Process automation, fraud/security, algo trading, robo-advisory

65
Q

Explain benefits of process automation

A

Replace manual work, automate repetitive tasks, and increase productivity

examples include chatbots and JPM’s COiN (Contract intelligence)

66
Q

Role of ML in security / fraud

A

Fraud detection, financial monitoring, underwriting and credit scoring

67
Q

Procedure of fraud detection

A
  1. Unsupervised learning (clustering)
  2. If any anomalies detected, trained AI model will separate legitimate and illegitimate transactions.
  3. Initial trianing of AI is using supervised learning
  4. Trained AI model is then given data, manual review by an expert, feedback (reinforcement learning) to become the final model
68
Q

Algo trading in ML

A
  1. Monitor trade results in real time and detect patterns
  2. Sentiment / news analysis
  3. Act to sell, hol and buy stocks
  4. Analyze thousands of data sources
  5. Squeeze slim advantage over market average → significant profits due to enormous volume
  6. Make thousand or million of trades - high-frequency trading
69
Q

Robo Advisory roles

A

Portofolio Management, and reccomendation of financial products

70
Q

SVM’s steps to do?

A
  1. Define hyperplane between classes with support vectors
  2. Optimise model to find support vectors that maximize the distance between hyperplane and classes (best line)
71
Q

What happens if there are 2 overlapping classes in SVM?

A

Use Kernel Trick - Map data to high dimensional space where they will be linearly separable

72
Q
A

FITE1010 (3 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions