AI

Question 1

What is a genetic algorithm?

Answer 1

• Begin with a population (a set of randomly generated possibilities)
• Rate each individual of the population based on some fitness function
• Create children from two randomly selected parents
• Mutate the children with some randomness
• Keep going until your children converge to best fitness

Question 2

What is an agent?

Answer 2

An agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators

Question 3

What is a percept?

Answer 3

the agent’s perceptual inputs at a given instant

Question 4

What is a percept sequence?

Answer 4

the complete history of everything the agent has ever perceived

Question 5

What is a rational agent?

Answer 5

chooses whichever action maximizes the expected value of the performance measure

Question 6

What is the difference between fully observable and partially observable?

Answer 6

Agent’s sensors give access to the complete environment or only some of the environment (relevant environment)

Question 7

What is the difference between deterministic and stochastic?

Answer 7

Deterministic - the next state is completely determined by the current state

Stochastic - too complex to be determined based on the previous state

Question 8

What are the different types of agents?

Answer 8

Simple reflex - makes actions based on the current percept and ignores the history

Model-based reflex - use a model of the world and its environment to make decisions

Goal-based - has model of the environment but also has a goal for desirable situations and/or outcomes

Utility-based - complex representation of a goal rather than just good or bad

Learning - agent is able to learn based on decisions, goals, and environment

Question 9

What is A* search?

Answer 9

BFS with a heuristic

Question 10

Basic NLP

Answer 10

Structure

• alphabet - set of characters to form words
• vocabulary - set of valid words in the language
• grammar - rules that connect words

Semantics
- meaning conveyed by the usage

Question 11

What is stop word removal?

Answer 11

Removal of commonly used connecting words (a, an, the, and, etc.)

Question 12

What is stemming?

Answer 12

Removing the endings of words to make the words more commonly used (study, studying, studied -> stem to study)

Question 13

What is lemmatization?

Answer 13

Create a common form of each word (running, ran -> lemmatize to run)

Question 14

Document-Term Matrix

Answer 14

Documents on the left, terms across the top

Question 15

Term-Document Matrix

Answer 15

Terms on the left, documents across the top

Question 16

Supervised learning

Answer 16

• Test and train set
• Use training set to find a particular observation, training set has attributes
• Used for classification

Question 17

Unsupervised learning

Answer 17

• Not looking for anything in particular, simply trying to find some kind of relationships between data

Example: k-means clustering

Question 18

What is clustering?

Answer 18

Grouping observations together based on some attribute or multiple attributes.

Question 19

What are decision trees?

Answer 19

Form of classification that makes binary decisions on hierarchical choices

Root is all observations
Make splits until everything is atomic
Leaves are actual outcomes

Question 20

What is neural networking?

Answer 20

Machine learning based on the human brain where each layer is a set of decisions designed to recognize patterns. Each layer provides positive or negative reinforcement toward the correctness of the classification.

Question 21

What is a support vector machine?

Answer 21

Used in combination with clustering, the SVM helps classify data with a vector of separation between clusters.