Artificial Intelligence, Machine Learning, and Deep Learning

9 min readJan 26, 2020

This article was written so that a grandmother can understand what is Machine Learning and Deep Learning.

What is Machine Learning?

It is a branch of artificial intelligence that began to gain importance from the 80s. It is a type of AI that no longer depends on rules and a programmer, but the computer can establish its own rules and learn by itself same.
Machine learning occurs through algorithms. An algorithm is nothing more than a series of ordered steps that are taken to perform a task.
The objective of machine learning is to create a model that allows us to solve a given task. The model is then trained using a large amount of data. The model learns from this data and is able to make predictions. Depending on the task you want to perform, it will be more appropriate to work with one algorithm or another.

What is artificial intelligence?

Artificial Inteligenics is the ability of a computer to learn by its own means to solve a specific problem. For example, whenever you pick up the phone at 7 in the morning, you open the FM radio application. Imagine that the cell phone was able to turn on the radio by itself every day at the same time without the need for you to do it?

That is artificial intelligence.

But that not only ends there, in artificial intelligence there are many different fields where, depending on how the instructions are given to the computer, it will respond. In art intelligence there are three different ways of learning for the computer through data:

Supervised Learning:

“In supervised learning, the algorithms work with” tagged “data, trying to find a function that, given the input variables, gives them the appropriate output tag. The algorithm is trained with a” historical “data and thus” learn “to assign the appropriate output label to a new value, that is, predict the output value. (Simeone, 2018)”

Unsupervised Learning

“Unsupervised learning takes place when there is no“ tagged ”data for training. We only know the input data, but there is no output data that corresponds to a given input. Therefore, we can only describe the structure of the data, to try to find some kind of organization that simplifies the analysis. Therefore, they are exploratory. “

Reinforced Learning

“This type of learning is based on improving the response of the model using a feedback process. The algorithm learns by observing the world around it. Its input information is the feedback or feedback it obtains from the outside world in response to its actions. therefore, the system learns from trial-error. “

For example, there is machine learning or supervised k-nearest neighbor:
It is a figure recognition model that can be used to classify and obtain graphic data on various objects.
Imagine we have the following figures to classify with this type of learning:

When new figures are added to the blank space, it is desired that the computer with supervised learning identifies the new figure and saves it in another place without deleting it from where it appeared:

if we tell the computer for example that it has a letter equal to 3, the machine will find the three closest figures of the new figure to classify them according to the shape of the largest number of figures it finds.
In the following figure, the three figures closest to the new figure that is the green heart are the dianamte, and two stars:

Among the simplest forms of supervised learning, the k-nearest neighbor method is considered a type of deferred learning, since generalization beyond training data does not occur until a system query is made.

Another supervised learning method is the Decision Tree Learning:

For general uses, this method is used to visually represent decisions and show or inform decision making.
The objective of learning the decision tree is to create a model that predicts the value of an objective based on input data.
When “learning” a tree, the source data is divided into subsets based on an attribute value test, which is repeated in each of the derived subsets recursively. Once the subset in a node has the value equivalent to the target value, the recursion process will be completed.

Let’s look at an example of several conditions that can determine if someone should go fishing or not. This includes weather conditions, as well as barometric pressure conditions.

In the simplified decision tree above, an example is classified by classifying it through the tree to the appropriate leaf node. This returns the classification associated with the particular leaf, which in this case is a Plaster to No. The tree classifies the conditions of a day based on whether it is suitable or not to go fishing.

The following learning model is Deep Learning:

Deep learning tries to simulate the abilities that your brain has when responding to the senses Granny, for example imagine that the agreement to the light it receives, the humidity that the earth has and the position in which your plants are could be pour the correct amount of water at the exact time of day? and that everything was controlled by means of a computer that was configured to understand what that exact moment is? It would be something great. Well, that is deep learning.

Now … how do you set up a computer and how do you tell the computer to do these kinds of tasks?

Neural networks and deep learning currently provide the best solutions to many problems in image recognition, voice recognition and natural language processing.

It is done through the programming languages:

It is a formal language that gives a person the ability to write a series of instructions or order sequences in the form of algorithms in order to control the physical or logical behavior of a computer.

What are algorithms ?:

It is a sequence of instructions that represent a solution model for certain types of problems or a set of instructions that are carried out in order to obtain the solution of a problem.

To program artificial intelligence the most used programming language is Python. It can be described as an open-source programming language, geared towards easy-to-learn objects because it is readable, powerful and direct both for preprocessing data and for working with data directly.

But quiet, to work with machine laerning with Python there are tools that are compatible with Python such as TensorFlow, PyTorch, Keras. The most used is Tensorflow

Because machine learning is used for the processing of large amounts of data or information, its most general applications are in financial services centers such as banks and other industrial companies, government agencies such as public security and public service companies. , the websites that recommend articles that might interest you based on the previous purchases you have made online, among many other applications.

One of the strongest branches in machine learning is deep learning.

Deep Learning or Deep Learning is part of a broader family of machine learning methods based on artificial neural networks. Learning can be supervised, semi-supervised or unsupervised.

Deep learning architectures such as deep neural networks, recurrent neural networks and convolutional neural networks have been applied to fields that include computer vision, speech recognition, natural language processing, audio recognition, social media filtering, automatic translation, bioinformatics, drug design, medical image analysis, material inspection and table games programs, in which comparable results have been obtained and, in some cases, superior to those of experts in the human field

Deep Learning or Deep Learning is part of a broader family of machine learning methods based on artificial neural networks. Learning can be supervised, semi-supervised or unsupervised.
Artificial neural networks (RNAs) were inspired by the nodes of information processing and distributed communication of biological systems. RNAs have several differences with respect to biological brains. Specifically, neural networks tend to be static and symbolic, while the biological brain of most living organisms is dynamic (plastic) and analog.

What is Deep learning?

As I had explained before in the 3 paths that have machine learning, Deep learning is a class of machine learning algorithms that uses multiple layers of neural networks to progressively extract higher-level features from the input data.

Everything sounds too interesting right? Now, what do I need to start working with machine learning and all its derivatives?

The main thing is to be clear that you should know about statistics, probqabilities, informatics, and mathematics, in general linear algebra. Despite the immense possibilities of Machine and Deep Learning, a thorough mathematical understanding of many of these techniques is necessary to fully understand the internal functioning of the algorithms and obtain good results.

Why Worry About The Maths?

There are many reasons why the mathematics of Machine Learning is important and I will highlight some of them below:

Selecting the right algorithm which includes giving considerations to accuracy, training time, model complexity, number of parameters and number of features.
Choosing parameter settings and validation strategies.
Identifying underfitting and overfitting by understanding the Bias-Variance tradeoff.
Estimating the right confidence interval and uncertainty.

What level of math do I need?

This graph could explain more clearly what level you should have if you want to work with machine learning:

Linear Algebra: A colleague, Skyler Speakman, recently said that “Linear Algebra is the mathematics of the 21st century” and I totally agree with the statement. In ML, Linear Algebra comes up everywhere. Topics such as Principal Component Analysis (PCA), Singular Value Decomposition (SVD), Eigendecomposition of a matrix, LU Decomposition, QR Decomposition/Factorization, Symmetric Matrices, Orthogonalization & Orthonormalization, Matrix Operations, Projections, Eigenvalues & Eigenvectors, Vector Spaces and Norms are needed for understanding the optimization methods used for machine learning. The amazing thing about Linear Algebra is that there are so many online resources. I have always said that the traditional classroom is dying because of the vast amount of resources available on the internet. My favorite Linear Algebra course is the one offered by MIT Courseware (Prof. Gilbert Strang).

2. Probability Theory and Statistics: Machine Learning and Statistics aren’t very different fields. Actually, someone recently defined Machine Learning as ‘doing statistics on a Mac’. Some of the fundamental Statistical and Probability Theory needed for ML are Combinatorics, Probability Rules & Axioms, Bayes’ Theorem, Random Variables, Variance and Expectation, Conditional and Joint Distributions, Standard Distributions (Bernoulli, Binomial, Multinomial, Uniform and Gaussian), Moment Generating Functions, Maximum Likelihood Estimation (MLE), Prior and Posterior, Maximum a Posteriori Estimation (MAP) and Sampling Methods.

3. Multivariate Calculus: Some of the necessary topics include Differential and Integral Calculus, Partial Derivatives, Vector-Values Functions, Directional Gradient, Hessian, Jacobian, Laplacian and Lagragian Distribution.

4. Algorithms and Complex Optimizations: This is important for understanding the computational efficiency and scalability of our Machine Learning Algorithm and for exploiting sparsity in our datasets. Knowledge of data structures (Binary Trees, Hashing, Heap, Stack etc), Dynamic Programming, Randomized & Sublinear Algorithm, Graphs, Gradient/Stochastic Descents and Primal-Dual methods are needed.

5. Others: This comprises of other Math topics not covered in the four major areas described above. They include Real and Complex Analysis (Sets and Sequences, Topology, Metric Spaces, Single-Valued and Continuous Functions, Limits, Cauchy Kernel, Fourier Transforms), Information Theory (Entropy, Information Gain), Function Spaces and Manifolds.