2017 is the foundation year for artificial intelligence development

AI has crept into every corner of our lives, but the technological revolution has just started.

According to EngadgetIt is now time to clarify that AI is a very broad term for a variety of technologies. We have artificial intelligence in virtual assistants like Siri, Alexa, Cortana and Google Assistant. You'll also find artificial intelligence in software such as Facebook's chatbot messenger, or Gmail's auto-reply feature. It is defined as "intelligence displayed by machines," but it also implies computer situations that can work without the guidance of the human person. Then there is the machine, when the computer teaches itself how to perform tasks that people usually do. For example, a facial recognition system at the Massachusetts Institute of Technology (MIT) has recently learned to recognize people as the way people recognize each other without the help of their owners. .

It is important that we do not confuse these concepts because machine learning is only part of artificial intelligence. We will use the term machine learning to talk about concepts such as neural networks and models such as Google's TensorFlow, and AI to refer to programs, devices, and software that perform tasks that they learned.

AI has surpassed humans in games like poker or go this year (photo: Engadget).

This year, the AI ​​has become too smart, like beating people at the poker or Go, to get maximum score in the game. Pac Man (which no human has ever done), and even equates to the most senior Super Smash Bros. players. We have also started using AI in medicine to predict disease and disease conditions, as well as detect users intent on suicide on social networks. Not only that, AI can compose music and write scripts.

No matter where you look, there is someone who is trying to add AI to something, and they are facilitated by neurons made by Google, Microsoft and many big companies. other investment this year. Machine learning has also grown rapidly, and it will continue to improve in the coming years.

One of the biggest developments as we prepare for 2018 is the transition from running cloud-based machine learning models to phones. This year, Google, Facebook and Apple have released framework versions (a complete library of built-in classes, a framework for developing application software) on their mobile devices. , which allows developers to accelerate AI-based tasks in their applications.

Chipmakers have also accelerated the design of processors for machine learning. Huawei, Apple and Qualcomm all have adjusted their latest chips this year to better handle AI-related tasks by providing dedicated neurons.

Basically, AI has been continually improving over the years, but mostly they rely on cloud technology. Take the image recognition system as an example. At first, it could distinguish between men and women, who looked very different. But as the program continues to "teach" more images in the cloud, it will be able to better identify, and those improvements will be sent to your phone. In 2018, we will put AI's processing power into pocket-sized devices. Being able to do these tasks right on your phone not only speeds up the AI, but it also helps protect your privacy when data is not sent to cloud servers.

Clearly, the industry is trying to lay the groundwork for smartphones and other devices to learn and improve themselves, such as interpreting, image recognition, etc. But when the amount of hardware is available The ability to handle AI tasks is increasing, developers are still having difficulty figuring out the best way to add AI to their applications. No one in the industry knows how to fully exploit AI, and at the same time resonates loudly.

Technology Animoji on iPhone X

As time goes by, every industry and every aspect of our lives - from supermarket shopping to auto rides - will be transformed through AI. Stores will know our taste, size and habits, then serve us or let us know where to go to get what we need. When you walk in, the retailer will know (through face or phone recognition) who you are, what you have bought in the past, what you are allergic to, the recent visit to your doctor, or the color you love. Artificial intelligence of the system will find out the tendency to buy your goods in a certain period of time, from which to make recommendations to more competitive products.

Cars will be smart enough to avoid obstacles along the way, using machine learning to better identify hazards and to calculate preventive measures. Even physicians will soon be relying on artificial intelligence to categorize radiographs, magnetic resonance imaging, and other medical imaging, and reduce the time it takes to diagnose the disease.

AI is already very popular in image recognition, and it will soon become even more popular. Security cameras are increasingly better at distinguishing people from people, animals or vehicles. Do not be surprised if one day they are used in legal judgments, criminal searches or missing persons.

Virtual assistants that we speak on the phone and smart speakers will not only be faster and talk more naturally by learning from conversations they will be able to better identify our needs. . For example, when you walk into your home after work, the lights go on, the temperature controller starts working, and your favorite music is played.

Of course, these things already exist, but the methods of implementation depend only on the settings you have installed. Artificial Intelligence will know how to adjust everything in your home the way you want, even when there are external factors that affect it, such as automatically turning on the air conditioner when you notice a change in temperature. outside. All these automations will turn the world we always dreamed of into reality (or the world of Wall-E, which is up to us).

In the years to come, we will see the AI ​​faster, stronger and really capable of changing our way of life. Maybe it will not happen right away, but sooner or later AI will control our lives - whether you want it or not.