WEBVTT

00:00:00.000 --> 00:00:05.080
Machine Learning is a branch of Artificial
Intelligence that enables computers

00:00:05.080 --> 00:00:10.880
to learn from data and perform tasks
that require human intelligence.

00:00:10.960 --> 00:00:12.880
Machine learning algorithms

00:00:12.880 --> 00:00:17.840
can analyze patterns, make predictions
and improve over time.

00:00:17.920 --> 00:00:21.120
Machine learning is not a new concept.

00:00:21.200 --> 00:00:26.200
It dates back to 1959,
when Arthur Samuel created a program

00:00:26.200 --> 00:00:30.200
that could play checkers
and learn from its own experience.

00:00:30.280 --> 00:00:32.800
Since then, machine learning has evolved

00:00:32.800 --> 00:00:36.320
and expanded
to many fields and applications

00:00:36.400 --> 00:00:39.240
such as social media, e-commerce,

00:00:39.240 --> 00:00:42.240
streaming services and health care.

00:00:42.400 --> 00:00:46.640
Machine learning can be divided into three
main types

00:00:46.720 --> 00:00:52.000
supervised learning, unsupervised
learning and reinforcement learning.

00:00:52.080 --> 00:00:54.800
Supervised learning is when the algorithm learns

00:00:54.800 --> 00:00:58.520
from labeled data, such as spam filters.

00:00:58.600 --> 00:01:01.440
Unsupervised learning
is when the algorithm learns

00:01:01.440 --> 00:01:06.080
from unlabeled data,
such as clustering algorithms.

00:01:06.160 --> 00:01:07.400
Clustering is a

00:01:07.400 --> 00:01:10.400
technique that grouped similar data points
together

00:01:10.600 --> 00:01:14.280
based on some measure of similarity
or distance.

00:01:14.360 --> 00:01:18.000
For example, customers might be clustered
based on age,

00:01:18.160 --> 00:01:22.720
gender, purchase history and satisfaction
ratings.

00:01:22.800 --> 00:01:25.640
Reinforcement
learning is when the algorithm learns

00:01:25.640 --> 00:01:28.640
from its own actions and rewards,

00:01:28.640 --> 00:01:31.320
such as self-driving cars.

00:01:31.320 --> 00:01:35.640
The algorithm calculates a score
based on how well it has performed in action

00:01:35.640 --> 00:01:40.240
like steering, accelerating,
braking and similar.

00:01:40.320 --> 00:01:45.440
It tries to maximize its score over time
by improving its strategy.

00:01:45.520 --> 00:01:49.440
Machine learning is a powerful
and exciting technology

00:01:49.520 --> 00:01:53.720
that can potentially transform
many aspects of our lives.

00:01:53.800 --> 00:01:56.800
However,
it comes with challenges and risks

00:01:56.960 --> 00:02:01.840
such as data, privacy, bias
and ethical issues.

00:02:01.920 --> 00:02:07.720
Ethical issues refers to the moral
and social implications of machine learning

00:02:07.720 --> 00:02:14.320
such as its impact on human dignity,
autonomy, justice and welfare.

00:02:14.400 --> 00:02:16.840
Therefore, it is important to ensure

00:02:16.840 --> 00:02:21.240
that machine learning is aligned
with human values and principles

00:02:21.320 --> 00:02:26.560
such as fairness, transparency,
explainability, and inclusiveness.