Lecture 1: Overview of Machine Learning

Question 1

What is Learning

Answer 1

Cognitive Process of acquiring knowledge

Question 2

What is Knowledge?

Answer 2

Information presented in the form of facts, patterns, concepts, rules, models, and skills

Question 3

What is Recognition?

Answer 3

matching sensory information or patterns with the existing knowledge in memory

Question 4

What is Understanding?

Answer 4

perception of intended meaning of the sensory information by integrating it with the existing knowledge to make sense out of it

Question 5

What is communication?

Answer 5

Exchanging of information

Question 6

What is Reasoning?

Answer 6

Reaching a conclusion through generalization, logical, and statistical approaches

Question 7

What is Planning?

Answer 7

Process of thinking about the activities to achieve a goal

Question 8

What is decision making?

Answer 8

Selection of a course of action among several alternative options

Question 9

What is problem solving?

Answer 9

finding solutions

Question 10

What is Imagination?

Answer 10

Thinking or creating something that may not exist based on the existing knowledge

Question 11

What is Learning?

Answer 11

Acquiring new knowledge and skills

Question 12

Where is knowledge stored?

Answer 12

1. Sensory Register (1-2 seconds)
2. Short-term/Working memory (20-60 seconds)
3. Long term memory (can be indefinite)

Question 13

What is Memorization?

Answer 13

Weak learning, computer implementation is trivial but still important

Question 14

What is finding patterns?

Answer 14

Identifying repeated forms, finding associations, and relationships

Question 15

What is Categorization and Classification?

Answer 15

Establishing abstract concepts from examples

Question 16

What is an Analogy?

Answer 16

Transferring information from one form to another for the purpose of explanation or clarification

Question 17

What is synthesizing?

Answer 17

combining results or knowledge (obtained from different learning methods) into a coherent whole for advanced or higher-level knowledge

Question 18

Definition of Machine Learning

Answer 18

A study of methods to develop a system that can learn on its own, without being explicitly programmed

Question 19

What kind of data is used for Machine Learning?

Answer 19

Labeled of ‘training data’

Question 20

What is the goal of using training data?

Answer 20

to find patterns or relationships, classifications, or generalizations

Question 21

What is Rote Learning?

Answer 21

Memorization and simple matching?

Question 22

What is supervised Learning?

Answer 22

Learning by examples in a training data set through generalization of induction (this is called training). There are usually two forms of problems: regression and classification

Question 23

What is Unsupervised Learning?

Answer 23

Clustering, finding associations or features in any form of data without training

Question 24

What is semi-supervised learning?

Answer 24

Mixing Supervised learning and unsupervised learning.

Question 25

What is reinforcement learning?

Answer 25

Learning rules or actions through trial and error

Question 26

What are other forms of machine learning?

Answer 26

1. Deep Learning 2. Online/Deep Learning 3. Ensemble Learning

Question 27

What is the difference between a regression and classification function

Answer 27

Regression: if the learned function is used to predict a continuous value Classification: the learned classifier is used to determine a discrete value

Question 28

Facets of Regression Learning?

Answer 28

1. ordinary least square method 2.Gradient descent 3. Maximum likelihood estimate

Question 29

What are the facets of classification learning methods?

Answer 29

1. Logistic regression 2. K-nearest neighbors 3. Artificial Neural Network 4. Naive Bayes 5. Decision Trees, Random Forests 6. Support Vector Machines

Question 30

What is the goal of unsupervised machine learning?

Answer 30

Clustering: Clusters C with similar properties Association: groups with high occurrences together Dependencies: groups with high correlations or variations

Question 31

What is the goal of Semi-Supervised Learning?

Answer 31

Labeled Data is hard to get, unlabeled data isn't. Create training data with a small set of labeled data and a large set of labeled data.

Question 32

What is reinforcement learning?

Answer 32

goal-directed learning through a series of actions that result in consequences as rewards.

Question 33

Is reinforcement learning supervised or unsupervised learning?

Answer 33

Neither, it is both

Question 34

What is shallow learning?

Answer 34

machine learning that does not recognize complex patterns

Question 35

What is a convolution neural network (CNN)?

Answer 35

An artificial neural network with some convolutional layers and some other layers

Question 36

What is Deep Learning?

Answer 36

Attempting to mimic the human brain with many layers of processing to extract features from data

Question 37

What is online/adaptive learning?

Answer 37

Updating prediction models in real time without comprehensive training process for new data.

Question 38

What is Ensemble Learning?

Answer 38

Use of multiple learning algorithms to obtain better predictive performance.