.
The seed is planted!
Let's start from the beginning, when the idea came up. You may be thinking that the Deep Learning technique has recently flourished, so it would have started a few 20-30 years, but let me tell you it all started a few 78 years. Yes, you read it right, la historia del deep learningDeep learning, A subdiscipline of artificial intelligence, relies on artificial neural networks to analyze and process large volumes of data. This technique allows machines to learn patterns and perform complex tasks, such as speech recognition and computer vision. Its ability to continuously improve as more data is provided to it makes it a key tool in various industries, from health... a menudo se remonta a 1943 when Walter Pitts Y Warren McCulloch created a computer model that supported the neural networks of the human brain. They used a mixture of algorithms and arithmetic which they called “threshold logic” to imitate the thought procedure.
Since then, Deep Learning has evolved steadily, with only two significant breaks in its development. Both were linked to the infamous artificial intelligence winters.
The sprouting of seeds is visible!
During the cold war, when American scientists were trying to translate from Russian to English and some of the greatest mathematicians like Alan Turing (often known as the father of modern computing) who created the Turing test to test the intelligence of a machine. Frank Rosenblatt, a mathematician came up with the first model based on neural networks called Perceptron in the year 1958. Esto es equivalente al modelo de aprendizaje automático Regresión logística con una Loss functionThe loss function is a fundamental tool in machine learning that quantifies the discrepancy between model predictions and actual values. Its goal is to guide the training process by minimizing this difference, thus allowing the model to learn more effectively. There are different types of loss functions, such as mean square error and cross-entropy, each one suitable for different tasks and... ligeramente distinto.
Inspiration: biological neuron
It is clear from history that we are always inspired by nature and this case is no different. This is very inspired by the nature and biology of our brain. At that moment, they had a very basic understanding of the functioning of neurons in our brain. Let me first introduce you to the biological neuron.
If we touch the surface level of a biological neuron, then it is mainly composed of 3 parts, core, dendrites and axons. The signs / Electrical impulses are received by these dendrites connected to the nucleus where the nucleus itself performs some processing and finally sends a message in the form of an electrical signal to the rest of the neurons connected by means of axons. This is the simplest explanation of the functioning of a biological neuron, people studying biology would be aware of how enormously complex its structure is and exactly how it works.
Then, those mathematicians and scientists came up with a way to mathematically represent this biological neuron where there are n inputs for a body and each one has some weights, since all inputs may not be equally important in producing the output. This output is nothing more than applying a function after taking the sum of the products of these inputs and their respective weights. Since this idea of the perceptron is far from the complex reality of a biological neuron, we can say that it is vaguely inspired by biology.
It's a sapling now!
Now came the era where people asked why we couldn't create a network of connected neurons that was again inspired by the biological brains of living creatures like humans., monkeys, ants, etc., that simply have a structure of interconnected neurons. Many attempts were made since the decade of 1960, but this was successful in a pivotal post on 1986 by a group of mathematicians, one of which was Geoffrey Hinton (has phenomenal contributions in the field of machine learning and AI).
Then they came up with the idea of Backpropagation algorithm. Briefly, we can remember this algorithm as a chain differentiation rule. Esto no solo hizo factible el trainingTraining is a systematic process designed to improve skills, physical knowledge or abilities. It is applied in various areas, like sport, Education and professional development. An effective training program includes goal planning, regular practice and evaluation of progress. Adaptation to individual needs and motivation are key factors in achieving successful and sustainable results in any discipline.... de la Red Neural Artificial, but also created a AI hype where people talked about it all day and thought that in the next 10 years it would be feasible for a machine to think like a human.
Even though it created such a stir, faded in the decade of 1990 and this period came to be known as the AI Winter because people promoted it a lot, but the actual effect was marginal at the time. What do you think could be the reason? Before I reveal the reason behind this, I'd like you to give it a chance.
To think…
To think…
Well, here you have.
Powdery mold on the floor!
Even though mathematicians came up with this beautiful backpropagation algorithm, Due to the lack of computational power in the decade of 1990 and the lack of data, this exaggeration to end died after the US Department of Defense. UU. stopped funding for AI seeing the marginal impact over the years after being so hyped. Then, machine learning algorithms like SVM, Random Forest and GBDT evolved and became extremely popular with 1995 Y 2009.
Mature tree with flowers!
Although they all went to algorithms like SVM and everything, Geoffrey Hinton still believed that true intelligence would be achieved only through neural networks. Therefore for almost 20 years, In other words, of 1986 a 2006, worked on neural networks. And in 2006 se le ocurrió un trabajo fenomenal sobre el entrenamiento de una red neuronalNeural networks are computational models inspired by the functioning of the human brain. They use structures known as artificial neurons to process and learn from data. These networks are fundamental in the field of artificial intelligence, enabling significant advancements in tasks such as image recognition, Natural Language Processing and Time Series Prediction, among others. Their ability to learn complex patterns makes them powerful tools.. profunda. This is the beginning of the era known as Deep learning. This post by Geoffrey Hinton didn't get much popularity until 2012.
You might be wondering what made deep neural networks extremely popular in 2012. Then, in 2012, Stanford held a competition called ImageNet, one of the most difficult problems back then consisted of millions of images and the task was to identify the objects from the given image. I would like you to remember that in 2012, people had a huge amount of data, and at the same time, the calculation was very powerful compared to what was present in the decade of 1980. The deep neural network or Deep Learning for that matter outperformed all machine learning algorithms in this competition.
This was the time when big tech giants like Google, Microsoft, Facebook and others began to see the potential of deep learning and began to invest heavily in this technology..
Ripe tree with fruits!
Today, if I talk about Deep Learning use cases, it is feasible it is important that you know that some of the popular voice assistants like Google Assistant, Siri, Alexa works with deep learning. At the same time, Tesla's self-driving cars are made possible by advancements in deep learning. Apart of this, it also has its applications in the health sector. I firmly believe that there is still a lot of potential in Deep Learning that we would experience in the coming years..
These are some of my social profiles that you may want to visit:
LinkedIn: https://bit.ly/3ltBarT
Github: https://bit.ly/3rQAYoH
Half: https://bit.ly/3a66Jn1
The media shown in this post is not the property of DataPeaker and is used at the author's discretion.
