Cyber-speech, the University cifrovoy the Russian language| 04.07.2019
The School of digital economy, far Eastern Federal University (FEFU SCA) create a digital corpus of Russian language learning machines, neural networks, and the development of a synthetic individual based on a
artificial intelligence. The project is implemented in the Laboratory of machine learning SCE of the University on the basis of the master’s degree programme “Artificial intelligence and big data”. Electronic collection
applications for participation in the first phase of the work will be opened in September.
One of the first products on the results of the work will be digital Manager — synthetic identity based on artificial intelligence capable of maintaining complex dialogs with the user,
ask the right questions and finding non-obvious answers and the clock to solve service tasks. On the similar principle will be to build call centers, systems, languages training,
translators, various expert systems, control systems for complex machine arrangements.
“We have already started the development of a synthetic personality, in partnership with Sberbank, which led to the formulation of more global problems. A significant challenge has been the lack of qualitatively marked base
Russian language for neural network learning. We plan to answer it together with our technical partner on machine learning — the company “Ashmanov neural networks”. It will provide us the technology
digital marking material. The results of our joint work, we gradually will transfer for public use for all interested parties.” — Said Ilya Mirin, Director of the School
the digital economy of the University.
The expert explained that in fact we are talking about the preparation of the academic corpus of the Russian language, analogues of which in the global scale exist for English and French languages. The
the main step is to collect an audio body and mark it in a special way that is understandable to the machine. Collection of material will occur through a website and mobile app.
The primary stage of accumulation of the language material is planned to be completed in a year, and then begin digitizing.
In the first phase of the work will involve volunteers from among the students of the University. Then join professional linguists, and specialists in computational linguistics that will focus on quality
layout audio: break it into parts of speech, put down the accent, pauses, divided into dialogues and monologues, will lead a spoken phrase to exact match the written text, and texts
well-read from the sheet will be separated from the spoken naturally. Thus we have to solve a range of related tasks.
In the 2019-20 academic year, the main research direction of the master’s degree programme “Artificial intelligence and big data”, open SCE in FEFU with the support of
Corporate University of “Sberbank” and “Ashmanov neural Networks”, will be natural language processing. The program recruited linguists, linguists, experts in Russian language and Russian as
foreign Asian languages, as well as journalists and software engineers.
Neural networks, Digitization, Artificial intelligence
Journal: Journal IT-News, Subscription to magazines
