Presentation is loading. Please wait.

Presentation is loading. Please wait.

Automatic Switchboard Operator Luboš Šmídl, Tomáš Valenta Department of Cybernetics Faculty of Applied Sciences University of West Bohemia in Pilsen.

Similar presentations


Presentation on theme: "Automatic Switchboard Operator Luboš Šmídl, Tomáš Valenta Department of Cybernetics Faculty of Applied Sciences University of West Bohemia in Pilsen."— Presentation transcript:

1 Automatic Switchboard Operator Luboš Šmídl, Tomáš Valenta Department of Cybernetics Faculty of Applied Sciences University of West Bohemia in Pilsen

2 Contents A dialogue system The dialogue Automatic speech recognition and speech grammar The data Advanced features Maintenance-free running Experiences and future Interesting facts Other applications

3 Purpose Automatic Switchboard Operator is a voice application whose purpose is to answer phone calls and transfer callers to requested persons. The caller makes input preferably by voice and the system informs him by voice as well. A voice dialogue system Whole UWB in Pilsen First such a large application of its kind in CZ

4 A dialogue system Data: MySQL, Oracle, … Document server: PHP Dialogue controller: VoiceXML Interpreter Speech engines: ERIS by SpeechTech and Dpt. of Cybernetics Telephony: SIP or ISDN

5 The dialogue Experienced vs. newbies Shortcuts Call n-th number Called person specification First name and surname Titles and degrees Department and function Voice or DTMF input Smith → 76484#

6 Automatic Speech Recognition Methods LVCSR Isolated words Grammars person = ( [(salut function) | (function salut) | salut | function] [degrees] (([firstname] surname) | (surname firstname)) [degrees] [function | department] ) | function;

7 Speech grammar complexity Prof. Ing. Josef Psutka, CSc., boss of DCy 1. Josef Psutka 2. Engineer Psutka 3. Boss of the Department of Cybernetics 4. Mister Psutka, professor 5. Professor Psutka, the Department of Cybernetics 6. Psutka Josef 7. etc. 26, 042 acceptable utterances

8 The Data Visual data vs. Aural data Prof. Ing. Psutka professor engineer psutka Generating pronunciations Rules-based, for TTS vs. for ASR Tomáš Tomáš, Thomas, Tom Fields tagging Better grammar matching, faster DB search J(firstname) P(surname) D(department) F(function) T(degree)

9 Advanced features Web presentation Administration Rules for pronunciations Shortcuts or Direct numbers Callers’ rights Phonebook searching Monitoring Statistics

10 Maintenance-free running Windows services, daemons Task scheduler 1. Import data 2. Generate pronunciations 3. Generate and compile grammar 4. Optional sanitary restart

11 Experiences Running since 2008 Extended grammar accepting Hello, Please, Thank you I would like to talk to Optimizing prompts Application made general Future Using statistics for person/number selection More info about employees More features and speed for experienced users New technologies: better TTS and ASR

12 Interesting numbers 2,095 persons 2,322 telephones 35,566,194 utterances 2.5 hours – grammar compilation time

13 Other dialogue applications  Entrance exams Since June 2000 3,000–5,000 calls a year  Exams Web access alternative  Recent news reading RSS from www.idnes.czwww.idnes.cz Categories: general, sport, economics, …  ASR demo Users can test ASR capabilities Web interface, users log in, own set of utterances  and others…

14 Thank you for attention Do you have any questions?

15 VoiceXML Mark-up language based upon XML Main advantage Minimizes client/server communication (more interactions in a document) Hides low-lever implementation details from the programmer Enables better portability Designed for content providers, dialogue designers Separates user interface (VoiceXML) from program logic Easy for both simple and complex applications VoiceXML Interpreter (like web browser) Document getter  Document interpreter (dialogue controller)  I/O interface – speech engine: telephony, ASR and TTS units Two kinds of dialogue: forms and menus


Download ppt "Automatic Switchboard Operator Luboš Šmídl, Tomáš Valenta Department of Cybernetics Faculty of Applied Sciences University of West Bohemia in Pilsen."

Similar presentations


Ads by Google