Successfully reported this slideshow.
You’ve unlocked unlimited downloads on SlideShare!
LLC «Speech Platform», S2S Next Group
• According to «Automatic Speech Recognition Applications Market 2010-2013», the
world speech technologies market is estimated at $900 million. The market growth
is about 28% a year.
• Russian speech technologies market is not large. According to “STEL – Computer
Systems”, its size is about $10 million, and the market dynamics is 15-20%. The
“Speech Technologies Center” company is more optimistic – it estimates Russian
market at $25 million.
• However, despite the foregoing,
there is still no simple and
user-friendly tool for
effective creation of
The speech technologies market is developing rapidly, penetrating all the spheres of our
everyday life. The Russian market is lagging behind, but recently people in this country
also face speech interfaces quite often, mostly while interaction with the systems of
It may seem that the quality of such systems wholly depends on the quality of speech
synthesis and recognition. However, there is an equally important factor which
influences the effectiveness of speech systems – the quality of a dialog as a whole. It
depends on successful completion of the system’s work and satisfaction of the user.
In connection therewith the problem of dialog construction appears to be urgent. It is
not a trivial task as it might seem. While creating a dialog interface one should take into
account numerous rules and recommendations which are not obvious for developers
having no experience of work with speech technologies. Besides, they have to execute a
great amount of specific work including efforts to create a dialog and integrate it with
ASR and TTS systems, create in-house databases and knowledge bases, etc.
All that makes it necessary to create a universal and user-friendly instrument, available
for developers, for speech dialog interfaces construction.
• The Platform is integration of three large technologies: speech synthesis (TTS),
speech recognition (ASR) and dialog support. The Platform also includes the use of
other speech technologies (dictation, speaker authentication, etc.)
• The Platform includes in-house data- and knowledge bases, standard corpora,
dictionaries and gateways for interaction with external information resources and
• Novelty and the main advantage of the Platform consist in the presence of an in-
house dialog engine integrated with technologies of third-party ASR and TTS
• The Platform development and commercial projects creation are planned to be
simultaneous. Commercial solutions are created while development of technologies
and expansion of functionality.
• Commercialization of the Platform technologies is to be via creation of in-house
commercial applications as well as through licensing of outside developers.
Commercial positioning of the project has two main directions:
1. The Platform developers will create their own cluster of unique speech services and
• Telephony solutions
• Mobile applications
• Corporate speech services
• Speech interfaces for existing services
2. Through license system external developers will create interactive speech interfaces
to their own applications based on the Platform.
Prototypes of the following systems have been created or underway at the moment:
BusinessVox – system of automatic inbound calls routing based on innovative speech
synthesis and recognition technologies.
MedVox – speech service of inbound and outbound calls processing for medical facilities.
MedVox.Doc – speech application for filling in the patient’s electronic health record.
PhoneLine Manager – speech service optimizing phone line handling via systems of
callback and auto-informing.
LogVox – speech service of utility counter data collection.
BankVox – speech service providing the caller on the fly with information about the
nearest ATMs and bank branches.
The prototype of the dialog engine is created.
The system of automatic voice call routing OfficeVox is in operation at the “RTI-
Sitronics” corporate group.
The system of automatic medical appointment MedVox is in operation at the Medical
Center of Information and Analysis of Irkutsk region.
Prototypes of commercial systems in the following segments are developed:
• corporate telephony,
• banking sector,
• housing and public utilities sector.
State registration certificates for computer programs and databases belonging to
the Platform are received.
Received state registration certificates:
№ 2011618581 “Program of keyword-based text dialog construction”,
№ 2012613366 “Speech dialog platform”,
№ 2013617971 “Voice dialog system of medical appointment;
№ 2012620482 “Database of popular Russian surnames in male and female forms
according to SAMPA standard”,
№ 2012620510 “Database of materials for speech recognition and synthesis systems
testing and quality evaluation”,
№ 2012615028 “Automated workstation for operating with program of keyword-
based text dialog construction”.
Patent on the method of human-machine dialog creation using the system of
semantic tree construction. Russia, assumed date – April, 2013.
• Favorable cost offer
• Ready dialog and database libraries for different applications
• Simplicity and convenience of use due to the user-friendly interface
• Simple access to the Platform resources
• Integration with different speech synthesis and recognition systems
• Possibility of technology choice according to specific functional needs and business-
• High-quality elaboration of dialog patterns
• Natural dialog orientation
• Intellectual call processing and logs analysis
• High quality of dialog results
• Multitasking and possibility of being used in different fields
Ksenia Zasypkina, CEO
Bauman MSTU, “Information systems and technologies”. Has great experience as project director.
Field of interest: IT, computer linguistics, and development of projects in the of applied linguistics,
Alexander Klachin, Head of the Board
Economist-orientalist, the Institute of Asian and African Studies at Lomonosov MSU. Received the
Global Executive MBA degree in the Fuqua School of Business, Duke University. Has more than 10
year experience of work in the field of computer linguistics, IT, telecommunication.
Boris Lobanov, Academic adviser
Minsk Radiotechnical Institute. D.Eng. SC. thesis “Methods of automatic phonemic text-to-speech
synthesis”, the Institute of Electronics and Computer Science of the Academy of Sciences of Latvian
SSR. A member of the European Speech Communication Association, a scientific expert of the
European network “Language Technologies”, the head of Belarusian subcommittee of the
International Speech Communication Association.
Alexander Kharlamov, R&D Director
MPEI, “Nuclear and electric power stations”, D.Eng. SC. Lomonosov MSU. Institute of Higher
Nervous Activity and Neurophysiology of RAS, senior researcher; Scientific-Research Institute of IT
and Telecommunications “Informika”, leading researcher; Federal Institute of Education
Development, head of laboratory; LLC “Unikor microsistems”, head of department.
THANK YOU FOR YOUR ATTENTION