240 likes | 461 Views
Synthetic Agents that Speak and Listen. Talking with Highbrow Avatars on Your Cell Phone Prof. Matthew Nickerson, Southern Utah University. Automated Audio Tours . Audio cassette player. Analog. Audio CD Player. Digital audio player. Multimedia player. Digital.
E N D
Synthetic Agents that Speak and Listen Talking with Highbrow Avatars on Your Cell Phone Prof. Matthew Nickerson, Southern Utah University
Automated Audio Tours Audio cassette player Analog Audio CD Player Digital audio player Multimedia player Digital
Research issues • Frustration and complications • Player damage, loss, or theft. • Patron anxiety • Updates and changes • Outdoor venues can be problematic. • Patrons with limited mobility.
Automated Audio Tours Audio cassette player Analog Audio CD Player Digital audio player B Y O P Multimedia player Digital
Voice Extensible Markup Language VXML is an XML-based markup language designed specifically to implement interactive voice dialogs. Web Server VXML Digital Sound Content User Cell Phone Voice / Telephony Gateway
Historical photograph exhibit A gallery exhibit featuring historic photographs covering 100 years of theater history in Cedar City, Utah. 1900-2000
Benefits to developer • Low upfront costs, start slow • No check out/in, maintenance, personnel • Easily updated • Real-time usage statistics • Powerful evaluation tool
Benefits to users • Familiar device • No check-in or collateral required • Avoid hygiene concerns • BYOD
Platform? Work with a Vendor Do it yourself
Partner with a vendor Web Server VXML Digital Sound Content User Cell Phone Voice / Telephony Gateway
User Cell Phone Bridging Worldwide Networks TELEPHONY INTERNET Voice Server Web Server - VXML Digital Sound Content
Built in VXML tools • Voice or DTMF input • Prerecorded or computer generated output • Audio system event handlers • Interrupt • Capture audio input
Virtual conversation A gallery exhibit featuring historic photographs covering 100 years of local theater history, 1900-2000.
Building Synthetic Agents • Voice or DTMF input • Prerecorded or computer generated output • Audio system event handlers • Interrupt • Capture audio input
Do you want to know more about General Lee? What artistic period are you interested in? What area are you currently exploring? Limiting Response Options • Ask questions
Limiting Response Options • Ask questions • Create grammars <rule id = “destination” scope = “public” > <one-of> <item> <tag> “new york” </tag> new york </item> <item> <tag> “new york” </tag> new york city </item> <item> <tag> “new york” </tag> big apple </item> </one-of> </rule>
PROXIMITY GEOGRAPHY SUBJECT Limiting Response Options • Ask questions • Create grammars • Point of contact
Location, location, location WiFi, GPS, Bluetooth
Challenges to Cultural Heritage Applications… and others • Current policies • Photography • Limiting phone calls/conversations • No speaker phones, please! • Reception
Choosing a voice Battle of the sexes among synthetic agents and avatars BMW, Unisys, GMVoices
Modulated human voice… ? Some swear that synthetic agents are better… others just swear. Clifford Nass, Stanford University; Sprint PCS
Nuance in “virtual” conversation Affective interpretation of metaphorical utterances Catherine Smith, et al. School of Computer Science, University of Birmingham