Rise of the Machines

PHOTO by Ramil Sitdikov / RIA Novosti
How artificial intelligence took over Russia’s Skolkovo innovation centre

The Skolkovo Centre's first Artificial Intelligence conference SKOLKOVO.AI was held on Nov. 14, 2016, in the innovation centre’s Hypercube building. Experts met to discuss not only the present and the future of the budding technology, but also took a trip down memory lane. Lenta.ru provides the highlights of this conference.

Many experts believe that artificial intelligence is the most important invention of humanity in the whole 100,000 years of its conscious existence. However, the meaning of the term has evolved over time. In the mid-20th century “artificial intelligence” generally referred to machine translation. While the automated creation of a more or less legible translation may seem ordinary today, back then it was something unachievable. Though American engineers managed to somewhat successfully translate German into English in the late 19th century, the technology’s heyday was only to arrive during the next one.

Interestingly, the scientific language of choice back then was German – machine translation was an efficient way to produce texts in English. Things changed after World War II and the fall of Nazi Germany. Russian became the dominant language in the scientific worlds. It should be noted that of the 400,000 American scientists, only 400 were fluent in Russian.

So American engineers decided to repeat their success and create new computers for automatic translation, this time from Russian into English. The first milestone was the IBM 701 – the giant computer, which was housed in several rooms, had a vocabulary of 250 words and several rules, with which it managed to translate 60 phrases from Russian into English. This achievement went down in history as the Georgetown – IBM experiment and was a huge success at the time. The US government thought that they were onto a winner and expected engineers to master Russian-English machine translation in ten years. They were wrong.

In 1966, young mathematician Alexander Kuleshov, now a member of the Russian Academy of Sciences and rector of the Skolkovo Science and Technology Institute, read the Pentagon’s open report, stating that the experiment with automatic translation ultimately failed, and that further methods were no longer being pursued. And so the only contemporary application of artificial intelligence was shut down.

Sar
An advanced anthropomorphic space robot SAR-401 on display at the 2016 Open Innovations Forum at the Skolkovo Technopark

New Horizons

Five decades passed. After half a century, artificial intelligence, in one form or another, is an integral part of our daily lives and its applications grow by the day. Experts who gathered at the inaugural SKOLKOVO.AI conference believe that AI is the future. At the same time, the discussion about what exactly is worthy of being called ‘artificial intelligence’ was constantly running throughout the event.

Everyone agreed with one thing: Kuleshov in his opening speech emphasised that AI, unlike any other branch of science or technology, has developed with many successes, but many failures as well – sometimes it was seen as the next big thing, sometimes it was dismissed as a fad. Up until the mid-2000’s the term “artificial intelligence” was shunned by the scientific community. They even called the period the “AI winter”. The AI hiatus was brought to an end by the discovery of deep learning algorithms, which paved the way for numerous applications.

“I believe today we are witnessing the spring solstice: the AI winter is over, summer is a long way to go, but spring is in full swing,” Kuleshov concluded.

Aiming for the Future

Skolkovo’s leading robotics expert Albert Efimov pointed out that interest in AI applications was growing rapidly. In Q1 of 2016, venture investments into AI startups in their early stages exceeded $1 billion; at the same time, the total investment volume since 2011 was estimated at $33 billion.

The development of AI poses certain global risks, though. Efimov noted that there are currently 230 jobs requiring cognition (accounting for 27% or $9 trillion, of global income), which are all at risk of being replaced by software. The spokesman believes that in a decade, the impact of AI on the job market will be in the range of $6 to $7 trillion.

A notable example of computers being better than humans at certain cognitive tasks is the product of Skolkovo’s own VisionLabs. The company created facial recognition software based on computer vision and deep learning. VisionLabs LUNA allows banks and retailers to identify and verify customers’ faces. This software has already saved Russian banks millions of roubles.

Skolkovo
Promobot information robot at the exhibition in the framework of Open Innovations - 2016 Forum in Skolkovo

Feeling the Way Forward

Another Skolkovo resident, Promobot, has recently presented its third generation of service robots at the recent Innovation expo. It signed contracts for 35 robots in just half an hour. The company manufactures robots which provide consultations, showcase promotional materials, boost sales via other methods, assist in navigation and collect relevant contextual data on the people they communicate with.

Igor Bogachev, Skolkovo’s vice president and head of its IT cluster, called the company’s success a “result of the hard work of Skolkovo’s complete ecosystem.” He also noted importance of the conference, praising the fact that it had over 1500 participants.

One third of them were Skolkovo startups and residents; 31% were corporate clients, investors and representatives of Russia’s development institutes and 27% were from the world of science and academics. 22% of attendees had an academic title and 22% had two or more degrees. These are the kinds of people which AI needs to grow as an industry, Bogachev noted.

Skolk

Sooner than You Would Imagine

Despite the fact that most amazing AI applications are still reserved for way off in the future, there are still short term forecasts which can be made. Albert Efimov believes that over the coming decades virtual assistants will become commonplace – perhaps they will be something of a status symbol. AI call centres will also become widespread – while they do exist now, their efficiency is, to say the least, sub-par. “If you can put a smart chip somewhere, they’ll do it,” the expert said. At the same time, this poses a threat, as there will be “a market of smart devices and a market for those who want to abuse the system,” Efimov warned.

There are more AI applications than one might expect. For example, there’s an AI-brewed beer in the UK – supposedly, it bases the taste of the product on customer reviews and requests. There is also an anti-rodent AI system which aims to curb diseases spread by mice and rats.

“Interest in the subject is patently obvious. The wave will die down, though, so the question is, will we ride it while we can? I believe that teams working in the field of artificial intelligence should seize the moment and start working. When the waves recede, we’ll see who was swimming without trunks,” Efimov concluded.