AVIOS Bounces Back
Defying reports of its demise, AVIOS held its 18th annual meeting in San Jose, Calif., May 24-26. Rumors of the end of the AVIOS conference were further put to rest by the size and quality of this year's conference. There were 249 paid conference attendees in addition to 150 others who visited the exhibit hall only. The keynotes and papers presented in the sessions clearly supported AVIOS' identity as a conference for speech-application developers. Many of the talks offered practical information about tools, techniques, testing, and usability. In addition, the exhibit hall was filled with products designed to make application development faster and easier. The conference included presentations covering a range of subjects, including toolkits, microphones, and embedded technology.
AVIOS has a tradition of including usability sessions and keynotes in its conferences. AVIOS' 99 was no different. Dr. Clifford Nass of Stanford University and co-author of The Media Equation presented the first keynote of the conference: How People Treat Computers, Televisions, and New Media like Real People and Places. Dr. Nass described some of the work he is doing to expose social rules that people apply to human-machine interaction. From that point on, the topic of usability flowed through the conference. Melissa Dougherty of Nuance Communications and Daryle Gardner-Bonneau of Michigan State University described user responses to specific application conditions and introduced ideas for improving usability. Kate Dobroth of the American Institutes for Research provided protocol for usability testing and a team of researchers from IBM offered a coding scheme for describing user behaviors. There were also presentations that addressed overall usability design, specific techniques, and other facets of usability.
Several of the usability sessions focused on designing dialog and "mixed-initiative" systems. These talks reflect a major shift in telephony applications from structured, menu-oriented systems towards more "conversational" interaction. A number of these sessions addressed the question of how much structure, if any, to include in these systems. In the Featured Address, "Evaluating Spoken Language Systems," Candace Kamm described an approach used at AT&T Labs to compare and evaluate dialog, mixed-initiative, and fixed-initiative systems using both subjective and objective measures. The measures were combined to produce an overall assessment of user satisfaction.
Bits and Pieces
The AVIOS' 99 conference and exhibition also recognized the emerging consumer-electronics and embedded-systems market. Chip-level products were prominently displayed in the exhibit hall and there was an entire session devoted to embedded systems. There was also a session describing Pitney Bowes' Universal Access Copier System. The copier can be operated using separable delegates, including voice and touch screen. The other end of the spectrum was represented as well. The conference included talks by SpeechWorks, Nuance, UNISYS, InterVoice, and others describing various approaches to component and/or modular design. Several modular and component-based development tools were also available in the exhibit hall. One
There was a session on microphone technology and one session on speaker verification and speech synthesis, but like most speech-processing industry conferences, the bulk of the AVIOS ' 99 conference was devoted to application development for speech recognition. For speech-recognition application developers AVIOS '99 conference offered a broad spectrum of high-quality sessions addressing developer-related issues and concerns.
Judith Markowitz is a leading independent analyst in the speech-recognition and voice-biometrics industries. She recently completed a market analysis of speaker verification and identification. She can be reached at 773-769-9243 or JMarkowitz@PObox.com