Fundamentals of AI

Question 1

What is Machine Learning?

Answer 1

A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E.

Question 2

What is Top Down?

Answer 2

Model all different functions and wire all the ‘agents’ together, deduction

Question 3

What is Bottom Up?

Answer 3

Give the system a lot of data, so it can discover the concepts itself, induction

Question 4

Three Pillars of Machine Learning

Answer 4

Models and Algorithms, Powerful and cheaper computation, Massive data warehouse

Question 5

What is Data Mining?

Answer 5

Exploration and analysis of large quantities of data discover valid, novel, useful and understandable patterns in data.

Question 6

What is Supervised Learning?

Answer 6

Infers a function from labelled training data, each example consisting of input and outputs (Classification and Regression)

Question 7

What is Unsupervised Learning?

Answer 7

Infers a function to describe hidden structure from unlabelled data (Clustering and Association)

Question 8

What can be done in Data pre-processing?

Answer 8

Fill in missing data, find outliers, feature selection.

Question 9

Unsupervised Learning - Clustering

Answer 9

Given: Un-labelled data set and similarities/distance metric
Goal: Find ‘natural’ partitioning, or groups of similar data points

Question 10

K-means clustering

Answer 10

Choose the number of k clusters and initialise K cluster centroids randomly. Assign each data point to the nearest centroid (based on distance), and update the centroids by the mean of all the data points assigned. Output the final cluster assignments and centroids.

Question 11

Application of k-clustering

Answer 11

Anomaly detection, Social Media Analysis

Question 12

Unsupervised Learning - Association

Answer 12

Discover correlation between any two or more variable.
Given: a set of records containing items
Goal: Produce dependency rules to predict occurrence of variable X with variable Y

Question 13

Categorical data

Answer 13

Learn to predict to which set an instance belongs to based on pre-labeled (classified) instances (Classification)

Question 14

Continuous data

Answer 14

Finds a linear relationship with the variable X and the variable Y

Question 15

Supervised Learning: Regression

Answer 15

Based om the given data find the function that minimises its mean squared error to fit the samples

Question 16

Overfitting

Answer 16

Describes errors in the dataset instead of the underlying relationship of the variables

Question 17

Regression Pros and Cons

Answer 17

Pros: short training time, easy to implement, easy to interpret
Cons: sensitive to noises and outliers (overfitting), cannot handle complicated relationships (linear only)

Question 18

Supervised Learning: Decision Tree

Answer 18

Internal nodes: decision rules on features
Branched: course of decision or action
Leaf nodes: a predicted class label (output)
Iteratively partition the decision space of chosen features.

Question 19

Decision Tree Pros and Cons

Answer 19

Pros: Reasonable training time, Caan handle large number of features, easy to implement, easy to interpret
Cons: only simple boundary decisions, problems with missing data, cannot handle complicated relationships, over-complex tree (overfitting)

Question 20

Neural Networks Pros and Cons

Answer 20

Pros: can learn more complicated class boundaries, can be more accurate, can handle large number of features
Cons: hard to implement: trial and error for choosing parameters and network structure, slow training time, can overfit the data, hard to interpret

Question 21

Supervised Learning: Neural Networks

Answer 21

Set of neurons connected by directed, weighted edges
Positive weight encourages the neuron to fire, while negative prevents firing. Each neuron is fixed at threshold t.

Question 22

Linearly Seperable

Answer 22

Where the output data pointscan be seperated using a linear boundary. Only a linearly separable function can be represented by a perceptron.

Question 23

Bayes Rule

Answer 23

Fundamental notion is of conditional probability

Question 24

KRR

Answer 24

Knowledge Representation and Reasoning

Question 25

Knowledge Based Systems

Answer 25

A system build around a knowledge base, i.e collection of knowledge taken from a human and stored in such a way that the system can reason with it