MDG Innovations

2017

Grant for 270 million rubles for development of technology of the analysis of the sound environment

On December 7, 2017 the MDG Innovations company received from "National Technology Initiative" (NTI) a grant for 270 million rubles. These means will go for development of technology of the analysis of the sound environment using the artificial intelligence (AI), the Vedomosti newspaper with reference to representatives of MDG Innovations and the state program writes.

It is about technology under the name "Neyroukho" which allows devices to distinguish sounds and in the presence of uncharacteristic noise (for example, at an excessive rumble at the enterprise) to send notices or to manage video cameras.

The Russian developer of biometrics MDG Innovations attracted 270 million rubles from the state on a sound analysis system using AI

Besides, developers promise to teach a system to define gender and age on a voice of the person. According to the partner of KPMG auditing company Nikolay Legkodimov, other similar systems have no such function.

The cost of all project is 390 million rubles. NTI selected money on condition of 30 percent private joint financing. 120 million rubles in technology will be invested by "The center of speech technologies" (CST; owner MDG Innovations). It is supposed that development will be able to pay off by 2022. And in two years the company expects revenue in 2 billion rubles.

According to the head of department of development of voice technologies and products of "Yandex" Denis Filippov, money there will be enough NTI on creation of a prototype, and on industrial implementation of technology additional resources will be required.

The managing director of Skolkovo Ventures Vladimir Sakovich says that state support can help such technologies as Neyroukho, to get quickly into public sector and on those markets where they will be demanded.

CST say that the voice biometrics can be useful in the field of the safe cities, robotics and "smart" houses, and this technology will warn workers of noisy productions about dangers.^[1]

2016

Received CST from IDF of 120 million on the system of face recognition

The MDG Innovations company from will gain from the Fund Industry Developments (FID) at Minpromtorg a loan in the amount of 122.4 million rubles. Means will go for creation of a hardware and software system of intellectual processing of video of CST: SmartTracker.

The total cost of the project will be 255.8 million rubles, i.e. the preferential loan of IDF will cover a half of expenses. A loan is issued due to the need of import substitution, its conditions correspond to the Development projects program. At the specified loan amount and the general project budget, it means charge of 5% per annum and repayment within five years.

At the same time the target sales volume of new products is not less than 50% of a loan amount a year, since second year of serial production.

The program of loans for 2.1 billion rubles

Providing company funds MDG Innovations are a part of the program of financing consisting of 9 concessional loans to projects on import substitution from IDF. In total the Expert advice of IDF selected 2.1 billion rubles for this program. At the same time the total cost of 9 approved projects exceeds 7.9 billion rubles the Fund claims that issue of these loans will allow to raise 5.8 billion additional resources to the real sector of economy.

Method of selection of coherent contexts of word usage

The MDG Innovations company (CST group) with assistance of the Ministry of Education and Science of the Russian Federation announced the method of selection of coherent contexts of word usage focused on work with the live unprepared (spontaneous) speech in April, 2016. The new method in such tools of speech analytics of CST as "the clustering module" and "the module of automatic annotation" will be used.

The method is based on technologies of work with Big Data (big data) and modern methods of machine learning (machine learning) in combination with deep linguistic and semantic analysis of the speech.

Especially sharply the problem of "understanding" costs the speech (the voice intelligence and/or large volume of speech data) in the large contact centers processing information in a constant duty and large volumes. For this reason the main potential consumers of the systems which are automatically selecting coherent contexts (logical structures, "situations") of these or those words, significant for the analysis, are analytical services of CC and technical support services of the government and commercial structures.

The purpose of selection of contexts in data bulks - to consider information which is directly connected with a target word and not to include in a context irrelevant information. Selection of coherent contexts of words many times simplifies work of analysts, allowing them to solve such problems in the automatic mode as:

• the analysis of the use of a key word (people, the organizations, geographical objects) in the text and/or the speech;
• identification of contexts of the use of a key (significant) word in dialogs;
• drawing up text summary (informative pressing) of the text and/or speech;
• automatic formation of headings of thematic clusters, etc.
  • 1 XLIFFService: Sequence contains no elements