[OSy] Projekt "Voice Interactions over Intelligent Networks"

[Petr Tuma]

Dobry den,

jak jsem jiz nekterym z vas rikal, objevila se moznost zucastnit se
projektu "Voice Interactions over Intelligent Networks" ve spolupraci s
tymem vyzkumniku z mistni vetve IBM T.J.Watson Research Center, kteri se
zabyvaji vyvojem aplikaci pro rozpoznavani reci.

Cilem projektu je posoudit perspektivu kombinace komponentoveho
middleware s aplikacemi pro rozpoznavani reci tak, aby bylo mozne
napriklad rozlozit vypocetni a komunikacni zatez spojenou s
rozpoznavanim reci (detaily viz dale).

Projekt by mel bezet velmi zhruba prvni polovinu roku 2006 a mel by se
na nem podilet relativne maly pocet lidi (vedle me a meho kolegy pana
Holuba tak 4-6 studentu). Pokud mate zajem, prosim kontaktujte mne co
nejdrive.

Diky, Petr Tuma

---<z popisu projektu>---

The project envisions using the principles of autonomic services to form
an intelligent network for real-time human-to-human and human-to-machine
interaction that would support multi-modal and sensoric services.
Relying on grid and peer-to-peer capabilities, the intelligent network
will host a distributed call center on voice over IP network with the
IBM speech recognition technology, with the goals of:

- improving resource utilization by migrating the components of the
distributed call center and the speech recognition technology. The
migration will optimize a cost function determined by factors such as
the character of the component, the network and node load, the user
location and interaction history, etc.

- improving management by employing the autonomic properties of the
components. The components of the distributed call center will be
instantiated and configured on demand to satisfy service requirements,
and reconfigured and upgraded as necessary.

The project will leverage the embedded speech recognition technology
developed by IBM Voice Technologies and Systems and the component
infrastructure developed by CU Distributed Systems Research Group to
provide a challenging and perspective platform.

The project would integrate additional applications, such as Skype or
SoftPhone for clients and Asterisk SoftPBX for servers, to provide the
other necessary components of the distributed call center. Due to size,
project is a feasibility study that delivers proof of concept prototype
in the following steps:

- assessment of legacy technologies, what can be split into components,
what can be reused.

- selection of target improvements in efficiency and reliability of the
intelligent network.

- deployment environment for the individual technologies.

- prototype with a static architecture.

- collecting information through monitoring.

- rules for infrastructure adaptation based on the output of monitoring.

During the course of the project, a number of students and research
staff will directly participate. Besides exposing the participants to
IBM technologies, the project will enable the participants to come in
contact with IBM employees.

-- 
Petr Tuma
Distributed Systems Research Group
Department of Software Engineering
Faculty of Mathematics and Physics
Charles University, Czech Republic
http://nenya.ms.mff.cuni.cz/~ceres