Alfanum: Innovations In Speech Technologies

Posted in: News- Nov 04, 2013

“Serbian Siri” – Talking With Your Smartphone

AlfaNum from Novi Sad, the regional leader in speech technologies for south-Slavic languages, organized the “AlfaNum Speech Tech Event 2013” on Oct. 11, 2013 in “Sava” convention center. The company displayed their latest speech technology-based products to an audience of more than 50 representatives of Serbian public enterprises and private companies (such as Telekom, SBB, VMA, OFPS, Komercijalna banka, etc).

Further improvement and commercialization of these technologies, which are being developed by a team of experts consisting of PhDs, experienced programmers and young researchers, is also supported by the Innovation Fund through its Matching Grants Program.

The item that stirred the most interest was “Voice Assistant”, a personal smart phone application based on AlfaNum’s complex technology for speech recognition and synthesis, as well as dialogue models which enable man-to-machine communication. This application was developed by AlfaNum’s spin-off company “Axon – Intelligent Systems”.

Navigating through a series of menus and options, dialing contacts and texting is all done by the user’s voice commands. This application, already nicknamed “Serbian Siri” after its famous Apple counterpart, will soon be available on Play Store. Also, there is a series of related applications designed for smart houses and different types of Interactive Voice Responses, with special detail being put into applications for people with disabilities.

“Voice Dial”, an application for direct and quick dialing of company staff (or outside contacts) using voice commands was also presented, along with “Advertising Monitor”, a powerful tool for searching through audio recordings, “Word Spotter” – an app designed for keyword-based searches in audio, uniquely built for the Serbian language, as well as “AudioMemo Registrofon”, a system for recording telephone conversations and other audio signals. This system enables precise, high-quality and long- lasting recording (with present hard-drive data storage capacities, it is possible to store several decades worth of recorded material).

Alongside the Serbian language text-to-speech synthesizer (which AlfaNum’s has already installed on numerous Serbian web pages), the company also organized a premiere display of speech synthesis for the Croatian language. This is an important part of AlfaNum’s intense regional development strategy, the company’s long-term goal. The main assets in AlfaNum’s future product offering are without a doubt the intelligent dialogue systems, which will take the entire man-to-machine interaction to a much higher level.

AlfaNum’s dynamic and successful team is led by CEO Darko Pekar and professor Vlado Delić, who as researcher and project manager at the Faculty of technical sciences, University of Novi Sad, created curricula in the field of acoustics, audio signal processing and speech technologies.