Artificial Intelligence, Machine Learning, and Deep Learning
( Complete Guide )
by Mahmut on January 08, 2021
When Google DeepMind’s AlphaGo program defeated South Korean Master Lee Se-dol in the board game Go, in October 2015, the terms AI, machine learning, and deep learning were used in the Media to explain how DeepMind won. And all three are part of the reason AlphaGo beat Lee Se-Dol. But they are not the same.
In this article, I am going to talk about artificial intelligence, machine learning, deep learning, and neural networks, and explain how they are different from each other.
In this article, you are going to learn :
What is Artificial Intelligence
We can define Artificial intelligence as the ability of a computer or machine to mimic human capabilities ( recognizing objects, understanding and responding to language, making decisions, solving problems ) without any human intervention.
After falling into science fiction for years, AI is a part of our daily lives today. The increase in artificial intelligence development has been made possible by the sudden availability of large amounts of data and the development and wide availability of computer systems that can process all this data faster and more accurately than humans.
Artificial intelligence complements our words as we write, gives directions when we ask, sweeps our floors, and then suggests what we should buy or watch repeatedly. It also drives applications such as medical image analysis that help skilled professionals do important jobs faster and more successfully.
The idea of ‘ a machine that thinks’ dates back to ancient Greece. Until today, AI has gone through many changes and developments. So let’s look at the developments in artificial intelligence until today.
History of Artificial Intelligence
Alan Turing published Computing Machinery and Intelligence. Turing, famous for breaking the ENIGMA code of the Nazis with the Second World War, said in the newspaper “Can machines think?” He proposes answering the question and introduces the Turing Test to determine whether a computer can show the same intelligence (or results of the same intelligence) as a human. The value of the Turing
test has been debated ever since.
John McCarthy coined the term ‘artificial intelligence’ at the first artificial intelligence conference at Dartmouth College. (McCarthy would continue to invent the Lisp language.) Later that year, Allen Newell, J.C. Shaw, and Herbert Simon created Logic Theorist, the first working AI software program.
Frank Rosenblatt developed the Mark 1 Perceptron, the first computer-based on a neural network that ‘learns’ through trial and error. Just a year later, Marvin Minsky and Seymour Papert published a book called Perceptrons, which became both a landmark study on neural networks and an argument against future neural network research projects, at least for a while.
Neural networks (algorithms for training the network) with backpropagation are widely used in AI applications.
IBM’s Deep Blue defeated world chess champion, Garry Kasparov, in a chess match (and a rematch).
Baidu’s Minwa supercomputer uses a special type of deep neural network called a convolutional neural network to describe and classify images with higher accuracy than the average person.
Backed by DeepMind’s deep neural network, the AlphaGo program beats world champion Go player Lee Sodol in a five-game match. As the game progresses (over 14.5 trillion after just four moves!) Victory is key given a large number of possible moves. Later, Google bought DeepMind for $ 400 million.
Artificial Intelligence Applications
Speech recognition, also called speech-to-text (STT), is an AI technology that recognizes spoken words and converts them into digital text. Speech recognition is a feature that runs computer dictation software, TV voice remote controls, voice text messaging, and GPS-enabled phone answering menus.
NLP enables a software application, computer, or machine to understand, interpret, and generate human text. NLP is the AI behind digital assistants (like Siri and Alexa mentioned above), chatbots, and other text-based virtual aids. Some NLP uses sentiment analysis to detect mood, attitude, or other subjective qualities in language.
AI technology that can identify and classify objects, people, text, and even actions in still or moving images. Typically used for image recognition driven by deep neural networks, fingerprint identification systems, mobile check payment applications, video and medical image analysis, self-driving cars, and much
Retail and entertainment sites use neural networks to suggest additional purchases or environments that are likely to appeal to the customer, depending on the customer’s past activity, other customers’ past activity,
and many other factors such as time of day and weather. Research has found that online recommendations can increase sales from 5% to 30%.
Designed to optimize stock portfolios, artificial intelligence-focused high-frequency trading platforms perform thousands or even millions of transactions per day without the need for human intervention.
Uber, Lyft, and other car-sharing services use artificial intelligence to match passengers with drivers to ensure reliable ETAs to minimize waiting times and diversions and even eliminate the need for excessive pricing during peak traffic periods.
It has been flying commercial and military aircraft for decades. Today, autopilot uses a combination of sensors, GPS technology, image recognition, collision avoidance technology, robotics, and natural language
processing to safely steer an aircraft in the sky and update human pilots as needed. Depending on who you ask, today’s commercial pilots spend as little as three and a half minutes manually flying a flight.
What is Machine Learning
Machine learning is a subset of artificial intelligence (AI) that learns from large amounts of data and focuses on creating applications that improve efficiency and accuracy over time without programming.
An algorithm in data science is a set of statistical processing steps. In machine learning, algorithms are ‘trained’ to find features and patterns in large amounts of data to make predictions based on new sets of data. Depending on how good the algorithm is, the more data is processed, the more accurate the decisions and predictions will be.
Today, we can see examples of machine learning everywhere. Digital assistants search the web and play music in response to our voice commands. Websites recommend products, movies, and songs based on what we have previously purchased, watched or listened to. Spam detectors prevent unsolicited emails from reaching your inbox. Medical image analysis systems help doctors detect tumors they may have overlooked. And the first driverless cars hit the road.
We can expect much more. As big data grows, computing becomes more powerful andaffordable, and data scientists continue to develop more capable algorithms, machine learning will provide greater efficiency in our personal and business lives.
Machine Learning Methods
Supervised machine learning uses tagged datasets to classify data or train algorithms to accurately predict results.
Supervised machine learning requires less training data than other machine learning methods and facilitates training because the results of the model can be compared with real labeled results. However, the process of preparing properly labeled data is expensive, and there is a danger of creating a model that is too closely tied and biased to training data that it does not properly handle variations in over-compliance or new data.
Unsupervised machine learning takes untagged data and extracts the necessary meaningful insights by using algorithms to tag, sort, and classify data in real-time without human intervention.
Unsupervised learning is about identifying patterns and relationships in data that people would miss, rather than automating decisions and predictions. Take spam detection, for example – people generate more email than a data scientist can hope to tag or classify in their lifetime.
Semi-supervised learning uses a smaller labeled data set to guide classification and feature extraction from a larger, unlabeled data set.
Semi-supervised learning can solve the problem of having not enough labeled data to train a supervised learning algorithm.
Machine Learning Applications
Apple Siri, Amazon Alexa, Google Assistant, and other digital assistants are powered by natural language processing (NLP), a machine learning application that allows computers to ‘understand’ human language the way humans do by processing text and voice data.
Deep learning models offer suggestions such as “people also liked” and “just for you”, as seen in Amazon, Netflix, Spotify, and other entertainment and retail services.
Chatbots can use a combination of pattern recognition, natural language processing, and deep neural networks to interpret the input text and provide the user with suitable responses.
Machine learning regression and classification models mark the use of stolen credit cards, replacing rule-based fraud detection systems that contain a high number of false positives and are rarely successful in detecting the criminal use of stolen or compromised financial data.
Almost every type of machine learning and deep learning algorithm mentioned above plays a role in enabling a self-driving car.
What is Deep Learning
Deep learning is a subset of machine learning in which multilayered neural networks modeled to function like the human brain ‘learn’ from huge amounts of data.
In each layer of the neural network, deep learning algorithms make calculations and gradually ‘learn’ by making predictions over and over, gradually improving the accuracy of the result over time.
Deep learning drives many artificial intelligence applications and services that improve automation, perform many analytical and physical operations.
Deep learning applications are part of our lives and are so well integrated into products and services that users are unaware of the complex processes that take place in the background.
So let’s look at the examples of deep learning applications.
Deep Learning Applications
Deep learning algorithms can analyze data to identify dangerous patterns that indicate possible criminal activity.
Speech recognition, computer vision, and other deep learning applications can increase the efficiency and effectiveness of investigative analysis by extracting patterns and evidence from audio and video recordings, images, and documents that help law enforcement more quickly and accurately analyze large amounts of data.
Financial institutions regularly use predictive analytics to assess business risks for loan approvals, detect fraud, and help manage loan and investment portfolios for clients.
The healthcare industry has benefited greatly from deep learning capabilities since the digitization of hospital records and images. Image recognition applications can support medical imaging professionals and radiologists, helping them analyze and evaluate more images in less time.
Differences Between Artificial Intelligence, Machine Learning, and Deep Learning
We talked about every one of them above. So Let’s talk about the differences between artificial intelligence, machine learning, and deep learning shortly one more time.
The easiest way to clarify the relationship between artificial intelligence (AI), machine learning, and deep learning is as follows:
In this article, we talked about Artificial Intelligence, Machine Learning, Deep Learning, and the relationship between them. If you have any questions or want to add something, please write in the comment section.