Vasily Mazin has been striving to improve AI-enabled systems and create a universal assistant for 10 years. He graduated from the applied mathematics faculty of Kuban State University where he held a position of a lecturer within 15 years. Mind Simulation is the result of his scientific work. At the company, Vasily is Director of AI Scientific Research.
In his interview to AI Conference, the expert shared the obstacles he faces while creating AI and explained his choice of hybrid systems.
“Every time I wanted to play a game on my PC, I had to write it first”
A personal computer that I received for my 14th birthday from the parents dramatically influenced my choice of the future profession. It was out of the ordinary Partner 01.01, like in the photo but red one. The computer didn’t have ROM, only a BASIC interpreter and an assembler. Every time I wanted to play a game on my PC, I had to write it first for the absence of memory.
Nonetheless, I liked the experience of game development, so I decided to enroll at the applied mathematics faculty. Before entering the university, I was studying at the physics and mathematics lyceum with a brilliant course of information technologies and programming.
“Being a junior, I wrote inventory management, accounting, business and financial administration programs that are available up to date”
I started working being a junior. I wrote inventory management, accounting, business and financial administration programs. Some of them are still available now.
As programs became more complicated, I headed the development team apart from programming. Then, we were focused on systems architecture and algorithms building.
“Term papers and theses gave an opportunity to check various ideas and approaches”
Having graduated from Kuban State University, I took over a position of a lecturer at the applied mathematics faculty. Teaching used to be a good part of the scientific activity. I had a lot of time after lectures to pursue science and related perform related tasks. Besides, term papers and theses allowed to check various ideas and approaches.
Later, I grew snowed under the academic activity and paperwork, so it was difficult to cope with other projects. As a result, I left the university two years ago dedicating my work to Mind Simulation.
“Already in 2003, the idea dawned upon me that neural networks have the right to life for both solutions to specialized tasks and parts of complex systems”
I took interest in AI and neural networks as soon as the first computer appeared in my life. At school and university, I created and programmed computer games (simulators of tanks and airplanes, strategies, role-playing games).
algorithms for rivals and game characters was the most interesting part of my work. I remember having written noughts and crosses on an unlimited grid where a winner had to put 5 crosses or noughts in a row. I created the game from scratch since both literature and the Internet weren’t available in Krasnodar at that time.
The program successfully won over freshers and amateurs. However, it was defeated by an experienced player.
I grew interested in neural networks in 2003 when they were part of my research activity. Then, everyone was shouting it is going to be a dead-end job. The idea dawned upon me that neural networks have the right to life for both solutions to specialized tasks and parts of complex systems. This led me to the concept of deep learning networks but there was no opportunity to practice it.
I am currently on the pulse of events in this industry. I have tested almost all existing AI approaches.
“We need AI not only to answer questions but also be an interlocutor, friend, a partner who understands you, maintains a conversation, and helps”
During my studying years at the university, I was thinking about the creation of a laboratory focused on AI tasks. In my free time, worked with a group of students but I didn’t manage to establish a lab at the university. So, as soon as the circumstances permitted, we created our own one.
At that moment, an interest in chatbots and voice assistants increased. Later, we got to know about Alice and scrutinized it. Then, we investigated Alexa, Siri, Cortana, and others.
It appeared that none of the abovementioned assistants could provide the essential thing – real communication. They call a taxi, show weather forecast, which is of benefit, but is that all? One cannot just have a talk with such assistants. We need AI not only to answer questions but also be an interlocutor, friend, a partner who understands you, maintains a conversation, and helps
Well-trained AI is the best solution in such cases. A single question came up: is this task worth meddling with? I got down to my previous ideas of knowledge representation and hybrid models. We estimated our possibilities and resources, drew up a detailed plan. That is how a new project aroused. We are currently focused only on it.
“Brainstorm of developers is a key to problem-solving”
Before getting started, we had general theoretical and partly algorithmic description of architecture and methods of AI core working. As soon we got down to its fulfilment, the difficulties of private nature related to a technical aspect of algorithms came up. Fortunately, brainstorm of developers was a key to problem-solving.
We faced a lot of difficulties with machine learning. Due to the lack of a critical knowledge mass needed for self-learning, we couldn’t automate the whole process at once. However, visible progress can be witnessed right now.
“We started to combine several methods more often, which led to the creation of powerful hybrids”
It took me many years to learn the application of methods in AI-related tasks and their role in problem-solving. During some experiments, we used two and more approaches simultaneously, which bore fruit. We started to combine several methods more often, which led to the creation of powerful hybrids.
“Knowledge and experience sharing is of benefit to science”
The audience shows a profound interest in AI-fueled products. Our presentation at Geek Picnic finished with a 40-minute interview session though we were the last to speak.
Knowledge and experience sharing is of benefit to science: you meet like-minded people, opponents, broaden your circle of acquaintances. This helps take another look at your work.
Of course, we currently cannot reveal all the details of our technology because it’s our competitive advantage. As time goes by, we will tell more and more information.
“We intend to open an AI school at the lab for the young generation to gain knowledge”
The possibility of development, discovering new directions, and product sales generates financial resources, that is why commercial success of Mind Simulation is important for us.
We intend to open an AI school at the lab for the young generation to gain knowledge helping them fulfill ambitions. This idea cannot be realized unless we collect the needed capital.
Before commercial distribution, we have to finish coding of basic core architecture. This project is divided into nine stages, we are currently finishing the fifth. The ninth will presumably be completed in February 2019. As of date, some organizations show their interest in using our core in the development of in-house products.
To meet Director of Scientific Research at Mind Simulation Lab, join us at AI Conference scheduled for November 22 in Moscow.