Rolling in the Deep Learning: Basic Concepts for Everyone
Alright Alright Alright! You probably know today deep learning is more relevant than ever. It is like synonymous with the modern era and lots of people love throwing this term around in conversation on social media. But why is it so? While the majority are just wowing, you can look deeper and find the answer.
Wanna catch a deep learning mechanism and finally know how everything works? If your answer is yes, then it is time to solve the mystery and know the magic behind it. In my attempts, I have done my best to help you with this matter and so I had showcased all important concepts regarding deep learning.
If you are a newcomer, don’t worry, cause I tried to avoid complicated terms, code pieces or math parts. It’s totally an easy and comprehensive manual for everyone, so grab your coffee and a comfortable chair and enjoy your learning.
Deep Learning, Machine Learning, and Artificial Intelligence — all together these cutting-edge technologies are connected, but not the same. For every curious beginner, it is better to start learning adventure with clarifying the differences between these buzzwords and see the whole picture.
In short, this picture has every element that comes out one after another. AI is an umbrella that contains everything: AI absorbs ML, and ML, in turn, absorbs DL. Here is a brief overview to figure it out better:
Artificial intelligence is a broad concept. At its core, AI is the presence of a wide variety of algorithms designed to simulate a human solution to the tasks. AI can relate to everything starting from software programs for playing chess and ending with speech recognition systems like Siri or Amazon Alexa voice assistant, which can perceive human speech and give answers to the questions.
Machine learning is an application of AI and application of the scientific study of algorithms and statistical models that computer systems use to perform a specific task without using explicit instructions, relying on patterns and inference instead. In simple words, ML is a technique by dint of which a computer does not just use a pre-written algorithm but learns how to solve the problem itself.
What is more, ML-application not only solves the tasks but also records the features of the solution for creating your own experience, simplifying and clarifying the next steps. Machines just receive data known as “training data” and “learn” from it. Then ML allows applying the knowledge gained during training on large data sets.
Unlike programs with manually encoded instructions for performing specific tasks, machine learning allows the system to learn independently recognize patterns and make predictions.
An example — a chess application studies the opponent’s behavior and takes it into account for the further reorganization of tactics.
Deep learning is a smaller element. It is a more specific application of AI, or more precisely a subset of machine learning and one of its algorithms. Right now DL is the most popular ML-algorithm that refers to a technique for creating an AI-powered layered neural network, much like a simplified replica of the human brain. DL gives computers the ability to solve more complex problems than other ML algorithms.
An example — during a game of chess, a neural network is trained predominantly.
Neural networks, like biological ones, are composed of neurons. In machines, the neurons are virtual — basically bits of code running statistical regressions. String enough of these virtual neurons together and you get a virtual neural network.
A neural network is a learning system, that is, it operates not only on the basis of given algorithms but also based on our own experience. A neuron in deep learning can be thought of as a “black box” with many inlets and one outlet. At the input, the neuron receives signals and forms an output based on them.
We know from biology that our ability to learn is based on the unique properties of a brain of 80 billion neurons.
