A socially oriented non-financial development institution and a major organizer of international conventions, congress, exhibitions, business, social and sporting, public, and cultural events.

The Roscongress Foundation is a socially oriented non-financial development institution and a major organizer of international conventions; exhibitions; and business, public, sporting, and cultural events. It was established in pursuance of a decision by the President of the Russian Federation.

The Foundation was established in 2007 with the aim of facilitating the development of Russia’s economic potential, promoting its national interests, and strengthening the country’s image. One of the roles of the Foundation is to comprehensively evaluate, analyse, and cover issues on the Russian and global economic agendas. It also offers administrative services, provides promotional support for business projects and attracting investment, and helps foster social entrepreneurship and charitable initiatives.

Each year, the Foundation’s events draw participants from 208 countries and territories, with more than 15,000 media representatives working on-site at Roscongress’ various venues. The Foundation benefits from analytical and professional expertise provided by 5000 people working in Russia and abroad. In addition, it works in close cooperation with 160 economic partners; industrialists’ and entrepreneurs’ unions; and financial, trade, and business associations from 75 countries worldwide.

The Roscongress Foundation has Telegram channels in Russian (t.me/Roscongress), English (t.me/RoscongressDirect), and Spanish (t.me/RoscongressEsp). Official website and Information and Analytical System of the Roscongress Foundation: roscongress.org.

RC personal account
Восстановление пароля
Введите адрес электронной почты или телефон, указанные при регистрации. Вам будет отправлена инструкция по восстановлению пароля.
Некорректный формат электронной почты или телефона
Art. Creativity. AI
4 June 2021
12:00—13:15
KEY CONCLUSIONS
What is art and what is creativity? Can AI generate AI?

I think the most important question is whether what we are watching or listening to triggers our emotions. You can make a movie based on algorithms, and this is something Hollywood is trying. <…> I understand that AI is now used a little to help with computer graphics. <…> If something triggers your emotions, it is art. If it does not, it is just a thing, — Ilya Naishuller, Director, Producer, Writer.

Goosebumps are what makes things created by people different form AI. <…> You will truly enjoy [the product created by AI, – Ed.] and you will learn something. Yet, no goosebumps. <…> It does not mean that you [developers, – Ed.] are unbale to trigger emotions – you will definitely trigger them. <…> I do apologize but it is not just about the inner processes – it is also about the inner light. I am a mathematician and I am telling you all neurophysiologists <…> claim that some things happen outside of the human mind. Some things are propelled by the unconscious. <…> These goosebumps are something a biological entity brings into its biological world, — Olga Uskova, Founder, President, Cognitive Technologies.

Art <…> is an elitist product. When we believe that we are art connoisseurs, we just know stories about art. We buy a painting not because we know something about it or because its colours tell us something. We buy it because <…> it was made by AI, which is cool and unusual. There is something new to it for the first or second time. So, we buy it. We buy a painting’s story, not the painting itself. The painting is attached to it, — Dmitriy Ozerkov, Director of Contemporary Art Departament, State Hermitage Museum, St.Petersburg, Director of «Hermitage 20/21», Project for Contemporary Art, Director of musicAeterna, Russia.

We keep forgetting certain professions – translators <…> for, example. I believe that it is a highly creative profession. You are not always able to translate something on the spot. If you translate things word by word, which we have known how to do for a while now, it looks weird. <…> When it comes to machine-generated translation, <…> what neural networks do is art of sorts, too. We are at the stage where neural networks – translations are done by neural networks – take a phrase in the original language and generate a phrase bearing the same meaning in a different language, — Alexander Kraynov, Director of AI Technology Development, Yandex.

It would be interesting if machines were able to come up with something humans cannot. <…> When it is not humans making up everything and machines doing the rest. Machines are generating whole situations. <…> It does not exist in art yet. <…> It should be something exciting, showy and new. <…> Argument, analysis and deconstructing the technology itself – this is what makes art interesting. So, a piece of art by AI needs to show that it has unravelled and overcome itself, — Dmitriy Ozerkov, Director of Contemporary Art Departament, State Hermitage Museum, St.Petersburg, Director of «Hermitage 20/21», Project for Contemporary Art, Director of musicAeterna, Russia.

AI is a tool for entertainment and real-life tasks

AI is basically a machine taught by humans that has a code. <…> It is a human-made tool. It means we get a product – a painting, a pot, or whatever – made with the help of a tool, — Dmitriy Ozerkov, Director of Contemporary Art Departament, State Hermitage Museum, St.Petersburg, Director of «Hermitage 20/21», Project for Contemporary Art, Director of musicAeterna, Russia.

Here is a practical task we have been on to: we needed to turn a small image into a big one. First, we did it for images and then applied it to movies. There is a technology called ‘super resolution’. It is based <…> on adversarial networks. The idea is pretty simple: we take a small image, and we need to make it big in such a way that the discriminator (the other network) could not tell the difference between this one and an originally big one, — Alexander Kraynov, Director of AI Technology Development, Yandex.

In a modern world, a personal assistant needs to have a sense of humour or knack for art. Our Alisa can paint, too. <…> Yet, it is not pragmatic yet – it is more of an entertainment, — Alexander Kraynov, Director of AI Technology Development, Yandex.

Google does not aim to create art with the help of AI yet. We have been using AI in a different fashion – it helps make art brighter and more exciting. It propels people to dive into art, — Julia Solovieva, Managing Director, Google Russia.

ISSUES
It is hard to find teachers who would teach AI to be creative

A computer as a screenwriter. <…> You know, there are very few good screenwriters. Who will teach Artificial Intelligence to write as the best screenwriters do? Tarantino does not even have a mobile phone, — Ilya Naishuller, Director, Producer, Writer.

No legal framework for copyright and intellectual property

As for intellectual property, it is regulated now. A person or an organization, who has intellectual rights, is considered the author. It is not like that in real life: one model is responsible for the second one. The latter one writes the third one. How we feel is not the only thing that matters. The legal aspect is also important. <…> Together with them [our colleagues from Rospatent, – Ed.], we are preparing an application for an intellectual product by Artificial Intelligence. It means that AI code was written by another AI code. So, we are applying for this AI product. It is not a film or a painting – it is a code. Yet, it is equally important. If together with Rospatent we are able to find the right procedure for intellectual property, this may become an actual breakthrough, — Alexander Vedyakhin, First Deputy Chairman of the Executive Board, Sberbank.

SOLUTIONS
Applying adversarial methods to AI learning

I have mentioned two adversarial networks. They compete and make each other better and more efficient. <…> You can use the basic concept of two adversarial networks to create images or generate styles that would remind you of certain artists. You can also use it to create artificial scientists. <…> In the future, [adversarial networks, – Ed.] will get improved, because our computing powers will be millions of times higher. <…> Most likely, we will get huge systems comparable to what happens in our brains. They will operate as artificial researchers and creators. They may exceed human researchers and artists. They may create amazing things, — Jurgen Schmidhuber, Chief Scientific Advisor, AIRI (Artificial Intelligence Research Institute).

Generative networks are absolutely wonderful. I generally believe that no matter what architecture you have, adding a discriminator at the end is always good. It improves the quality. If you have some sort of critic at the end, it is always beneficial, — Alexander Kraynov, Director of AI Technology Development, Yandex.

Goosebumps are <…> a matter of network learning and reinforcing it. These approaches sometimes generate more goosebumps and sometime less. Let’s address the approaches that generate more goosebumps. <…> Since we have a range of <…> products made by Artificial Intelligence, it makes sense to study them. Once we start studying them, we will be able to assess what gets better and what gets worse. It will end up in improvement eventually, — Alexander Vedyakhin, First Deputy Chairman of the Executive Board, Sberbank.

Importance of setting goals and objectives and clear terms of reference

Initially, it is not the Artificial Intelligence that writes the code for Artificial Intelligence. Terms of reference are the foundation for all AIs. This is what we develop. We will write something that will evolve step by step. <…> Yes, we generate something. Terms of reference cannot be underestimated. I keep saying everywhere I go: ‘Guys, do not strive for an all-purpose Artificial Intelligence – we do not need it. <…> It needs to be target-specific’. If we use AI in music as a tool, we need to understand what we are doing it for. If we use it for painting, we need to understand our purpose. It means that terms of reference are key, and this is something done by creators, — Olga Uskova, Founder, President, Cognitive Technologies.