Before looking ahead and seeing what tomorrow has in store, let’s take a step back to find out about the history of artificial intelligence and what the main reasons behind its creation are. The first steps that led to the birth of this discipline date back to the 1600s. It was only in the last century, however—in 1959 to be exact—that Arthur Samuel, a pioneer in artificial intelligence, suggested that, instead of receiving everything they needed from programmers, computers could learn autonomously. This belief was consolidated over the years, and with the spread of the Internet and the increased use of sensors and mobile devices, it became possible to create and aggregate huge amounts of data from which machines are now able to extract meaningful information.
In the world of artificial intelligence, machine learning (or automatic learning) represents a further step forward. This is because, thanks to specially designed applications, machines are able to learn autonomously by analyzing and comparing data. The role of humans is to define how the program should learn as well as choosing the examples and information that need to be stored and determining how the program should apply the knowledge in order to make decisions. Countless types of techniques are used for analysis, from the most basic ones, combining analytic geometry and linear algebra, up to very complex algorithms.
One of the most important techniques is deep learning. This technique employs neural networks, a mathematical model that simulates the behavior of a human brain, and may take the form of either software-based or hardware-based systems. By exploiting the potential of artificial neural networks based on multiple levels, connections and directions of data propagation, it is possible to carry out services and implement applications in a way that would have been unthinkable only a few years ago.
Nervana Systems and Intel, working together to improve deep learning
Without a doubt, one of the most innovative and promising companies in this industry is Nervana Systems. Recently acquired by Intel, Nervana has paved the way in creating optimized solutions to solve machine learning problems. In future, the time needed to train artificial intelligence will be greatly reduced, which will have the immediate consequence of improving the performance of machine learning and deep learning models.
Today, Nervana System is already working on three highly innovative technological solutions. The first is called Neon, which is a logical architecture used in the creation of models. The second is Nervana Cloud, a platform designed for business and for the creation of customized solutions. Finally, we have the Nervana Engine processor, developed with the aim of improving the performance of deep learning systems. So far, the processor has made deep learning up to 10 times faster compared to current solutions.
AlphaGo, the computer that challenges humans
AlphaGo reached an important milestone in the field of artificial intelligence, in the form of a piece of software for the game Go designed by the DeepMind research team. AlphaGo challenged one of the best Go players in the world, Lee Sedol, to a series of five matches, winning four and only losing one.
For the first time, man was defeated by a computer at one of the oldest and most complex games of all time. The technologies implemented in AlphaGo allowed the software to analyze every single move and to understand and anticipate Lee Sedol’s behavior. This is a clear demonstration of the potential of today’s technologies. Far from considering this episode as a victory of computers over humans, we believe the contrary: AlphaGo demonstrates the great strength that lies in combining human knowledge with the computational power of computers.
Artificial intelligence: everywhere, invisible and useful
The spread of machine learning and deep learning technologies is changing our everyday lives. Another interesting example—although it may seem trivial considering that we are now accustomed to using them constantly—is how the keypads on our smartphones work. Our keypads operate via a predictive system that provides suggestions based on context and on our writing style. Working via real time analysis, our phone keypad can simultaneously learn and predict the words that we are going to type. Even applications such as face recognition tools, music recognition apps and similar programs are still based on these automatic learning technologies.
As you can imagine, the use of artificial intelligence is expanding in every industry. Driverless smart cars, search engines, analysis models for the medical industry, predictive systems: these are only a few of the applications where AI is currently being used — applications that are constantly evolving.
AI, machine learning and deep learning: what will the future look like?
Analyzing the impact that artificial intelligence will have on our future is part of a wider debate, even in the scientific community. There are different schools of thought, one of which fears a future where artificial intelligence itself will manage to replace humans. This idea is heard frequently in today’s society, where science fiction movies often show how beings with higher cognitive abilities than those of humans could effectively take over humanity. In real life, however, it is more relevant to discover how these technologies will be used by humans. In fact, it is not so much the technology itself but rather the way in which it is going to be used that will make the difference. The improvements which are already being made to our everyday lives by these innovations are truly extraordinary.
What is certain is that we are witnessing a paradigm shift in the way we lead our lives. The potential is vast, and there is no doubt that what we are seeing today is only the first chapter in a story that will be written over the coming years.
By Alberto Marini