N-Tech.Lab is Pushing the Boundaries of Artificial Intelligence

Chances are you have seen some futuristic movie that addresses the multiple possibilities of artificial intelligence, such as the work of Steven Spielberg that rightly takes its name ("A.I. - Artificial Intelligence"). For years, humans have wondered how technology can advance the point of replicating characteristics that are typically human characteristics, and today we can say that this future is really close to become a reality.

And that happens mostly because of companies like N-Tech.Lab, who are pushing and taking artificial intelligence to whole new levels. Artem Kukharenko, a Russian developer fascinated by technology since early ages, is the face of this company, which specializes on facial recognition, using neural networks as the foundation for their work.

N-Tech.Lab became the noticeable player when it won the Megaface face recognition challenge, a well-known competition held by the University of Washington, beating even Google's Facenet algorithm. Participants were asked to develop an algorithm able to find the exact person among one million people using only one photo and N-Tech.Lab's algorithm demonstrated the highest accuracy.

It is now capable of finding a specific person among the photos of one billion of people, in less than one second.
Since then the team has developed the algorithm even further and it is now capable of finding a specific person among the photos of one billion of people, in less than one second.

N-Tech.Lab became known to almost everyone when Findface emerged, a face-recognition project based on their platform. Findface allows users to find similar looking people in the biggest (over 350 million users) social network of Eastern Europe, VK, which is basically the Russian Facebook created by Pavel Durov, the man behind Telegram, another buzz-making app. Findface has received over a million downloads and signups during the first months, with no marketing promotions, due to the viral effects.

After he graduated, Kukharenko abandoned facial recognition for three years, and moved his focus on neural networks and machine learning. He traveled to South America, while he was working on a freelance basis for the laboratory of Purdue University. Artem was working on the algorithms that allow to classify objects presented on video.

They worked on the algorithm to use neural networks in car auto-pilot systems for automatic classification of objects on the road, such as buildings, pedestrians, and road signs. After he returned to Russia, Kukharenko joined the Russian division of Samsung and continued to work on neural networks.

In the beginning of 2015, along with his girlfriend, he decided to put his knowledge in facial recognition to good use, he mapped faces of dogs and developed an app that identified dogs by taking a picture of them.

Magic Dog, as the app was called, only scored around 10,000 downloads, with user feedback pointing some bugs, such as identifying a cat as a dog. One feedback in particular stroke Kukharenko's mind: to expand the app to identify people. He did so, and decided to present the app to investors and the Russian venture fund "Typhon Digital Development".

The result of those negotiations is N-Tech.Lab. Kukharenko eventually got one fourth of its shares. Although he suggested a variety of tasks that could be solved with neural networks, the company decided to focus just on facial recognition. The team already developed an algorithm, "FaceN", which operates using the neural network that is capable of learning distinguishing face details useful for personal identification, such as eye size, eyebrow thickness, lip shape, and so on. Kukharenko explains further:

"We train the neural network with millions of mapped photographs. In a semi-automatic mode, people in the photos are identified as John, or Jack, or Stephen. Then the network learns by itself, trying to extract the vectors of features that would solve the task."

FaceN generates about 80 numbers to describe all the information about a face and, funnily enough, the team is still trying to understand what several of them mean. After entering and winning a facial recognition contest held in the US, where their product actually beat one made by Google, the offers to buy their algorithm came like a flood.

N-Tech.Lab got its first investment within just a couple of months, and then Kukharenko quit Samsung, but also took one of his fellow programmers with him. Another developer in the team was found in a VK community dedicated to neural networks, simply by browsing through comments.

In the future N-Tech.Lab is planning to launch the cloud face recognition software platform, which will be available for every business to plug into and use for their own recognition tasks. As they claim, the platform can be used for a great variety of areas, including security purposes, like ID checks at and realtime CCTV analysis, retail solutions for targeted advertising to in-store customer, dating services, entertainment, for instance casinos can use the Big Data processing technologies or amusement parks which deliver the photos right to the people caught on them.

One of the most interesting offers came from the Government of Turkey, who wanted to apply it to identify people crossing their borders, but they also had other offers coming from intelligence services, from Russia and abroad. Last May, N-Tech.Lab announced an agreement with the city of Moscow to test the facial-recognition service on the city's CCTV camera network. Kukharenko explains how this will work:

"People who pass by the cameras are verified against the connected database of criminals or missing people. If the system signals a high level of likeness, a warning is sent to a police officer near the location. There is no such system in any other city around the world."

Yet another potential use of FaceN is the 2018 FIFA World Cup, which will be held in Russia, as it could be used as a way to detect banned soccer fans. Regardless of the applications FaceN might have, some legal questions arise: is it legal (or ethical, to say the least) to use such a system? Is this system invading people's privacy?

Those questions will surely have a response in a near future but, for now, it is probably better to focus on the amazing potential that this technology has, and all the applications it might have in other distinct fields.