The current artificial intelligence market continues to take its first steps according to the data. And we presented in early 2020 from researchers at the Artificial Intelligence Observatory 2020.
If we dig deeper, we see that intelligent computing initiatives are the primary AI project type, accounting for 33% of total spending. This is followed by development projects for Natural Language Processing and Chatbot / Virtual Assistant, making up 28% of the market. In terms of sectors, four have turned out to be more than others.
What we have been living in since the 80s is a time of great fermentation in the fields of artificial intelligence. We can compare it to that of a person in the prime of his youth, full of energy and experimentation. But with the appropriate ability to apply what has learned, analyze, evaluate its potential and possible developments. Here, our metaphor ends because the next step that of maturity is still on the horizon. But here are the critical unions of the past 40 years divided by strands.
Table of Contents
Neural Networks
We’ll get into this in detail later when we talk about deep learning, but here are some basic historical terms.
In the above para, we saw that the first artificial neural network theory (the Rosenblatt perceptron) had challenged, but in 1986 Jay McClelland and David Rumelhart published parallel distributed processing.
Explorations into the microstructure of knowledge that lay the foundations of connectionism. Conclusion and studies in this area give new strength. So a neural network is a nonlinear direct graph in which each processing element (each node of the system) receives signals from other nodes and sends an alert to other nodes.
In practice, knowledge in a computer concentrated in a particular place, memory. The ability cannot localize in the neural network. It distributed in the structure’s connections so that the system can learn from its own experience.
The multilayer neural network model (MLP-Multilayer Perceptron) developed in which each node layer fully connected to the next. And uses a supervised learning technique called error backpropagation (which we will explain later in the in-depth learning chapter) for network training and Agile Training goes by one result that is already known to a particular champion.
Artificial Intelligence Support Nanotechnologies
The development of nanotechnologies, i.e., the ever-increasing miniaturization of microprocessors, has led to developing a new generation of components (from general-purpose GPUs) ResistiveRAM to anthropomorphic chips) that have given artificial intelligence enormous computing power thanks to the original impetus available. We will see them in detail later in the chapter on Technologies that enable and support Artificial Intelligence.
Innovative Algorithms
Suppose the previous topics (with a final parallel to our metaphor) represent artificial intelligence’s physicality. In that case, the algorithms should mean what comes closest to thinking. And this “magical” act makes people very special compared to other essence makes animals.
It is precisely from the development of innovative methods and algorithms that artificial intelligence draws its soul. In them, we will focus on the parts of this article. And devoted to machine learning, deep learning, natural language processing, and augmented reality, and virtual reality.
Artificial Intelligence Support Cognitive Computing
Use of self-learning algorithms, data mining, and big data analysis, pattern recognition, natural language processing, signal processing (a signal is a temporary variation in the physical state of a system or a physical quantity that used to represent and transmit messages or information from the far, so signal analysis is a component that supports cognitive computing). Implementing the most advanced hardware technologies creates technological platforms that try to mimic the human brain, from more specific activities to increasingly complex processing.
From the historic IBM Watson. So the first commercial supercomputer of its kind, to Google Deepmind and Baidu Minwa, several examples are now available.
