Definitions Flashcards Preview

Machine Learning > Definitions > Flashcards

Flashcards in Definitions Deck (15)
Loading flashcards...
1

What is a decision boundary?

 a hyperplane that partitions the underlying vector space into two classes

The classifier will classify all the points on one side of the decision boundary as belonging to one class and all those on the other side as belonging to the other class.

2

What is Two-point crossover?

In two-point crossover, two crossover points are picked randomly from the parent chromosomes. The bits in between the two points are swapped between the parent organisms.

Two-point crossover is equivalent to performing two single-point crossovers with different crossover points. This strategy can be generalized to k-point crossover for any positive integer k, picking k crossover points.

 

 

3

What is Intelligence?

  • (human) Intelligence is the ability to understand and learn things.
  • intelligence isthe ability to think and understand instead of doing things by instinct or automatically.

4

What is Hebb's law?

If two neurons on either side of a connection are activated synchronously, then the weight of that connection is increased.

If two neurons on either side of a connection are activated asynchronously, then the weight of that connection is decreased

5

What is soft computing?

an emerging approach to building hybrid intelligent systems capable of reasoning and learning in an uncertain and imprecise environment.

The combination of probabilistic reasoning, fuzzy logic, neural networks and evolutionary computation forms the core of soft computing,

6

What is a hybrid approach?

A hybrid approach is one that combines at least two methods. For example, combining a neural network with a fuzzy system results in a hybrid neuro-fuzzy system

7

What is Diversity?

Networks belonging to an ensemble are thought to be diverse with respect to a test set if they make different generalisation errors on that test set. Different patterns of generalisations can be produced when networks are trained either on different training sets, or from different initial conditions, or with different numbers or hidden nodes, or using different algorithms.

8

What is Knowledge?

Knowledge is a theoretical or practical
understanding of a subject or a domain.

9

What is a domain expert?

Those who possess knowledge in a domain are called experts.
Anyone can be considered a domain expert if he or she has deep knowledge (of both facts and rules) and strong practical experience in a particular domain.

The area of the domain may be limited.

In general, an expert is a skilful person who can do things other people cannot.

10

What is the purpose of Knowledge Representation?

Knowledge acquired from experts or induced
from a set of data must be represented in a
format that is both understandable by humans
and executable on computers

11

What is reasoning?

a mechanism for selecting
the relevant facts
and extracting conclusions from
them in a logical way.

We used it in the inference engine of KBS.

12

Elitism

Elitism is an optional characteristic of a GA that makes sure
that the fittest chromosome of a population of N
chromosomes is passed on to the next generation
unchanged; it can never be replaced by another
chromosome.

Without elitism this chromosome may be lost.
Extended forms of elitism are also possible where the best m
chromosomes of the population are retained.
Simple elitism is the case where m=1.
The effect of elitism is that the number of offspring that are
generated each generation is reduced from N to N-m
replacing the worst N-m individuals in the population

13

Ensebmle

Ensemble learning

• Ensemble based methods enable an increase
in generalisation performance by combining
several individual models trained on the same
task.
• Approach has been justified both theoretically
and empirically.
• The creation of an ensemble is often divided
into two steps : (i) generate individual ensemble
members and (ii) appropriately combine
individual members outputs to produce the
output of the ensemble.

14

Resilietn backprop?

The Rprop method : help to eliminate harmful influences of
derivatives' magnitude on the weight updates.
Basic Idea: the sign of the derivative is used to determine the
direction of the weight update; the magnitude of the
derivative has no effect on the weight update.

15