SpeechTEK has always provided an excellent overview of commercial speech in call centers, but one piece that was missing was hands-on experience with technology from a variety of vendors demonstrating comparable solutions. Last year, the SpeechTEK organizers corrected that omission by introducing SpeechTEK Labs. This year, they have expanded the labs to include a session on speaker identification and verification (SIV Lab). I’m in charge of the SIV Lab, which will be held on Monday, August 24, from 1:15–3 p.m. I would like you to participate.
The SIV Lab will offer hands-on experience with SIV enrollment and speaker verification (SV) using solutions from different vendors. Participants will see how vendors differ in both technology and their approaches to the same basic tasks. No sales materials or presentations are permitted.
In the SIV Lab, attendees will test the solutions and be asked to rate them on a small set of parameters, such as how well the technology/solution worked, how easy it was to use, and how much guidance was needed. I encourage you to comment on your experience with each vendor. You may also comment on the lab itself, so we can make it better next year. This feedback can be anonymous. Your assessments will also be given as feedback to the vendors. Each vendor will see only the evaluations for its own solutions.
Seven companies are participating in the lab, presenting their unique approaches to enrollment and verification. Each will first enroll you in its system and then have three “tries” to illustrate its verification and other solution(s). Those other solutions vary and might include adding a different kind of SIV or non-SIV authentication (e.g., knowledge verification) or performing speaker identification or speaker tracking.
Here are the confirmed SIV lab participants:
- Avaya, a leading provider of contact-center solutions, recently strengthened its position in the midsize contact center market through the acquisition of Agile Software. As an integrator, Avaya provides solutions using SV technology from other companies. Offerings include outbound as well as inbound solutions.
- Convergys is also an integrator. Its Intervoice division was one of the earliest developers of SV solutions for call centers. Today, its SV solutions support the company’s enterprise-level, relationship-management focus and include premises-based options, as well as the newly released On Demand Voice Authentication.
- Nuance Communications provides SV technology and solutions to customers and a network of integrators. The company commercialized three processes: applying SV and speech recognition to the same input, “variable-length authentication” (stops when sufficient data is collected), and knowledge verification (asking content questions).
- PerSay is the first vendor to commercialize text-independent SIV. It operates on free-flowing speech, but can also be used for text-dependent (password) and text-prompted solutions where the system tells the user what to say (e.g., Say “13672”). The company is probably best known for Bell Canada’s nationwide deployment to more than 800,000 customers.
- Recognition Technologies’ solution is fundamentally text-independent but can be used in text-dependent and text-prompted solutions. The company offers technology for SV, identification, classification, and tracking. I featured it in a recent Speech Technology magazine column (“Speech for Distance Learning,” January/February 2009).
- Speech Technology Center (STC) has text-dependent, text-prompted, and text-independent technology and solutions. In addition to offering call-center technology, STC is a market leader in speaker recognition for forensics and lawful intercept, and has been very active in the development of international standards for speech and SIV.
- VoiceVault pioneered the “combination lock” approach to challenge response (e.g., Say “45 76”) as well as hosted SV. The company has a strong focus on security. Its Dublin data center holds ISO 27001 certification, which is an international standard for information security management that prescribes basic standards for information security.
I also encourage you to attend SpeechTEK’s other SIV sessions on Monday at 10:15 a.m. and 11:15 a.m., and SpeechTEK’s other labs: Speech Synthesis (Monday, 10:15 a.m. to noon), Mobile Devices (Tuesday, 10:45 a.m. to 12:30 p.m.), The Voice-Enabled Enterprise (Tuesday, 1:45–3:30 p.m.), and In Beta, featuring products under development (Wednesday, 10:45 a.m. to 12:30 p.m.).
Judith Markowitz, Ph.D., is president of J. Markowitz Consultants and a leading independent analyst in the speech and voice biometrics field. She can be reached at firstname.lastname@example.org.