YouTube Transcript:
¿Qué es Machine Learning?

Skip watching entire videos - get the full transcript, search for keywords, and copy with one click.

Video Transcript

[Music]

although it is often confused with

other terms machine learning is a

discipline contained within

artificial intelligence, that is to say, it is

a soft set of

artificial intelligence that is also often confused

with deep learning which is in turn a

soft set, machine learning in

general terms of

artificial intelligence can be defined as the

use of computers or machines

to imitate the way in which the

human mind solves problems or makes

decisions machine learning is a

discipline of artificial intelligence

that focuses more on giving

machines the ability to learn from

data and past experiences through

a series of algorithms that feed

back so that

machines can identify patterns and

make predictions without

human intervention machine learning enables

computers to operate autonomously

without the need to explicitly program them

machine learning applications receive new data and

from it they can learn grow

and adapt and the performance of

machine learning algorithms improves

progressively as the

number of samples it receives increases during the

learning process

how a basic

machine learning algorithm works in the vast majority of

cases it all starts with

training data that this

training data is added to an algorithm

not without reading ning from that

a data input model is generated

then the algorithm itself appears and then

from the algorithm the

prediction appears then it will be determined how

accurate that prediction is if the

prediction is not accurate then

the algorithm is trained again with more

data and if it is accurate then

the model is considered successful and

from that data can be added, that is,

if the algorithm can predict with a

certain level of certainty then we can

start using that algorithm by

adding input, that is,

input data according to the model already defined and the

algorithm will be able to make

predictions from it

machine learning algorithms can

be trained in many

different ways each of which has

its pros and cons based on this and

the different ways in which

algorithms learn the chin

learning scheme is mainly categorized into

four types supervised unsupervised

semi-supervised and

reinforcement learning and rainforest meant learning

let's see them one by one in

supervised learning the machines are

trained using

perfectly categorized training data the term

in english for this type of data is

playboy this means that the data that is

used to training the algorithm are

perfectly categorized with the

correct output meaning the

training data given to the model

acts as the supervisor that teaches

the machines to predict the output

correctly the same concept applies

in which students learn under the

supervision of a teacher you could

say that super bikes learning is the

process of giving input data as

well as the correct output data

to the machine learning model the idea is

that the algorithm can make a

direct mapping between an input variable and

an output variable how

exactly does supervised learning work

these types of models are delivered using

categorized data sets where the mode

learns about each type of data

once training is

complete the model is checked against

test data which is a sa set of

the training data and then the

model predicts the output let's suppose we have

a dataset for different

shapes which include triangles

squares and circles the first step is to

train the model for the different

types of shapes if the shape has three

sides then it will be categorized as a

triangle if the shape has four sides

then it will be categorized as square

if the shape has a single side then it

will be categorized as circle after

training the model we are going to test it by

adding a test data set and

we are going to determine if the algorithm's prediction is

acceptable or not if it is not we

will go back to the first step and add

more categorized data in turn

supervised learning can be divided

into two types of algorithms regression and classification

classification

regression algorithms are used if there is a

direct relationship between the

input variable and the output variable it is generally used

for the prediction of

continuous variables such as

weather forecasting with market trends for

example some of the most

popular regression algorithms are

linear regression regression trees nonlinear regression

among others

classification algorithms on the other hand are used when

the output variable is categorical in

the sense that there are only two

classes such as or sign or

true or false some of the

most popular uses of these algorithms are

for example spam detection and

email filtering and some of the

best known algorithms are random forest

sad decision and logistic regression

as the name suggests

unsupervised learning is a

machine learning technique in which the models

are not supervised using

training data sets but

instead the model itself has to

finding patterns in the data you

receive can be compared to the way

learning occurs in the

human brain a more

technical definition would be unsupervised learning

is a type of machine

learning in which models are

trained using unclassified data sets

and are allowed to act on the data

data

supervision this type of model

cannot be directly applied to

classification or regression problems

like the ones we just saw because although we have

data the input we do not have the

corresponding output data

the objective of unsupervised learning

is to find what the

structure of the input data set is,

group them according to their similarities and

represent that data set in a

more compressed format that

can then be extrapolated how it works

exactly let's look at an example where

we have a data set that has not been

classified meaning that it has not

been categorized and the

corresponding output data has not

been provided either this data is

sent to the model to train it the

model will first interpret the

raw data to find

hidden patterns in the data set and then

apply the algorithm that it considers necessary

necessary

once the algorithm has been applied

this algorithm divides the data into

groups according to the similarities it

finds and the differences among objects

objects

the unsupervised learning algorithm

can be categorized into two

types of problems clustering, which groups

objects into clusters such that

objects with more similarities between

them remain in a group and have

less similarities with objects in other

groups, cluster analysis looks for

common elements in objects and

categorizes them according to the presence or absence

of these common elements and there

is also the association an

association rule is an unsupervised learning method

that is used to

find relationships between variables in

large databases a very common use

is marketing, for example, customers

who buy product x also

tend to buy the product and the

semi-supervised approach, as its name

indicates, is an intermediate point between

supervised and unsupervised learning

in this case, a

small amount of classified input data is combined

with a large amount of

unclassified data during the

training phase, in many cases getting

classified data is very complicated and

expensive, that is why only a

small amount is available. however, when

unclassified data is used in conjunction

with a small set of classified data, they

can produce a considerable improvement

in learning accuracy.

rainforest mente learning deserves a

dedicated video so I will not go into

too much detail here if this This is

content that interests you let me know

in the comments

reinforcement learning is the science of

decision making the idea is to find the

optimal behavior in a

given environment to obtain the maximum reward

reward

this optimal behavior is learned

through interaction with the environment

and observation of how the

environment responds equivalent to what

a child would do to explore the world

around them and learn what actions

lead them to achieve goals as there is no

training data or a

supervisor the model must independently discover

what sequence of

actions maximize the reward this

discovery process is basically

trial and error because this

model can learn from its own

actions in completely

new environments in explored is that it is considered

a very powerful algorithm

this would be the most common diagram in a

rainford tendler ning algorithm the

agent is the entity that executes the

actions in an environment in which it

is seeking to obtain a reward

once people interact with the

environment then it receives a new state

of the environment which allows it to learn and

if appropriate it will also receive a

reward let's suppose that we are

training a dog we cannot

directly tell it what we want because the

dog does not understand then we look for a

different strategy we do something and see

how the dog responds since it can

respond in different ways if the

dog does what we want it to do

then we give it a reward in this case

the dog is the agent that is exposed

to an environment that could be the house or the

park for example an example of the state

could be what the dog is doing, that

is, if it is sitting or what it is looking at

if the dog is sitting let's suppose we

use a specific word for it to

walk in people react and perform

an action that takes it from one state to

another the dog stops sitting and

starts walking

after the transition it may or may not

receive a reward or a punishment depending on the

result of the action this is a

classic example of reinforcement learning, that is,

reinforcement can be

positive or negative, positive reinforcement

increases the strength and frequency of

those behaviors that

positively affect the action taken by the

agent while negative reinforcement

is responsible for focusing on hardening

the behavior that occurs as a

consequence of a negative condition

that must be stopped or avoided in other

words, one focuses on the times

that people are right while the

other on the times that people are

wrong reinforcement learning is

used mainly in robotics and

driving autonomous some of the

most common uses of machine learning today are

for example the health industry

increasingly machine learning is being

adopted in the

health industry and surely you are even already

using it if you have any device

with sensors that measure your physical condition

such as smart watches or

any gadget that monitors your health

in real time additionally the technology is

helping medical teams analyze

certain trends based on

body metrics or detect events that

could help make certain

diagnoses or predict and prevent

future problems and even today

these algorithms are allowing

medical experts to predict the

remaining life span in patients suffering from

terminal illnesses for example another

very common use today is in the

financial sector today many

financial organizations use

machine learning to detect

fraudulent activities or also to

identify investment opportunities

for themselves and their clients the

reality is that today many

financial institutions are

hiring the services of

technology companies to squeeze the

benefits of machine learning

also in sales many

e-commerce sites use machine

learning to recommend items based

on the purchase history of

users these sites use

machine learning techniques that capture data

analyze it and deliver

personalized shopping experiences according to the user

sure is experienced this use case

yourself especially when it comes

to marketing campaigns also in the

transportation industry scared über

uber used a wide variety of

machine learning algorithms one of

them for example the one that sets the

price of the trip dynamically based on a

series of environmental factors such as the

schedule the availability of cars the

distance the initial location etc. it

used a

real-time prediction system to identify patterns in

traffic supply and demand etc. in

some places they are even

experimenting with driverless übers

that use the autonomous driving system

that we saw before in short

the use of machine learning is

increasingly adopted it is already being adopted we

could even say that it is one of the

technologies that are doing the most for

the evolution of civilization and what do you

think let me know in the

comments box and also if you liked this video

please leave me a like and if you have

n't done so yet of course subscribe to

the channel I have much more content

like this and much more content to

come to finish I want to tell you

that there is a platform that helps you

find your next job talent

click has offered me a discount on the

tuition for this program for all of you

who stayed here and have at

least 2 of experience in the

software industry you can find the link in

the first comment of this video click it

and tell him that you are from decoder

caip thank you very much for staying

here and see you next time [Music]

Click on any text or timestamp to jump to that moment in the video

Most transcripts ready in under 5 seconds

One-Click Copy125+ LanguagesSearch ContentJump to Timestamps

Paste YouTube URL

Enter any YouTube video link to get the full transcript

Most transcripts ready in under 5 seconds

Get Our Chrome Extension

Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.

Add to Chrome — Free

Works with YouTube, Coursera, Udemy and more educational platforms

Get Instant Transcripts: Just Edit the Domain in Your Address Bar!

YouTube

←

→

↻

https://www.youtube.com/watch?v=UF8uR6Z6KLc

YoutubeToText

←

→

↻

https://youtubetotext.net/watch?v=UF8uR6Z6KLc

YouTube Transcript:¿Qué es Machine Learning?

Video Transcript

Paste YouTube URL

Transcript Extraction Form

Get Our Chrome Extension

Get Instant Transcripts: Just Edit the Domain in Your Address Bar!

YouTube Transcript:
¿Qué es Machine Learning?