Over the past couple of years Google has been making big investments in cloud-based speech recognition services and applications, particularly for the Android and Chrome platforms. Dave Burke discusses and demos Google’s latest developments and touches on some of the company’s future plans in this area.

This session examines how speech technology can impact fundamental business decisions. How would your web application change if users could speak and listen to it? The first presentation describes how voice can be added to web browsers to achieve just that. Where should you host your IVR system: on premise or in the cloud? The second presentation discusses the advantages and disadvantages of various approaches to hosting the cloud-based computing.

This panel session will help you to benefit from the experiences of individuals who developed and deployed real applications for public transportation, medical documentation, and warehousing. Who is the target audience for the speech technology application and what services does the application provide? What problems did they encounter during implementation and deployment, and how did they overcome these problems? What measures of success did they use and what was their return on investment? What key lessons did they learn?

Two experienced analysts explore how speech recognition can be used effectively to improve customer interactions and how to leverage the technology to improve the overall customer experience.

In recent years driver distraction has become a growing problem. Solutions include using text-to-speech solutions to read messages aloud, and speech recognition to allow hands-free messaging. However, some feel these solutions do more harm than good by creating yet another distraction. With consumers becoming increasingly dependent on mobile devices, the problem is bound to grow. Can voice user interfaces be specially designed and tested for incar use to provide a safer way to interact by text whilst maintaining productivity on the move?

What are voice biometrics and how can you use them to verify who the caller really is? Discover the advantages and disadvantages and analyse how voice biometrics compares with other caller identification techniques. Learn how to determine which identification techniques callers will accept, and explore whether passwords and challenge dialogues will supplement or be replaced by biometrics? How much security is enough? 

Help us celebrate the opening of SpeechTEK Europe and meet your speech colleagues from around the world at our informal drinks reception.

Over the past couple of years Google has been making big investments in cloud-based speech recognition services and applications, particularly for the Android and Chrome platforms. Dave Burke discusses and demos Google’s latest developments and touches on some of the company’s future plans in this area.

The success of a natural language dialogue relies heavily on the correct balance of several factors, such as fallback menus and visibility of the operator.  Large scale ‘Wizard of Oz’ testing on live traffic in call centres can help in determining the right balance.  This approach gives statistically valid data about customer behaviour and language in advance of costly engineering and natural language tuning cycles. Example data showing the lessons learned will be drawn from realworld use in scores of tests from Europe and the United States in recent years.

The World Wide Web Consortium (W3C) is standardising several new languages to enable new functions and capabilities in your IVR systems: VoiceXML 3.0 (the new version of the popular language for developing speech applications); Call Control XML (language for controlling telephone connections); EmotionML (represents emotions captured from user behaviour and emotions exhibited in artificial agents); and HTML-Speech (embedding speech into web pages). Learn about the key features of these languages and when they will be available.

The first presentation describes why popular techniques for describing dialogues, such as finite state machine models, call-flows and form-based models, are fundamentally sub-optimal. Learn how to specify dialogues by describing WHAT rather than HOW using a purely Declarative Dialogue Model. The second presentation focusses on the challenges presented when IVR systems fail to help a user, resulting in call backs. Learn how to predict the reasons for call backs so you can refine the IVR system and provide better results the first time the user calls.

Prior to this conference, VUI experts met to identify and discuss design issues for cross-linguistic and cross-cultural voice interaction speech applications. The first presentation summarises their deliberations and presents guidelines for designing systems for use in multicultural and multilingual environments. The second presentation singles out the most unique student entries in the AVIOS student programming contest and suggests how we all might extend our vision to embrace some new directions.

These case studies reveal how IVR platforms provide advanced services to customers. The first presentation describes a platform which leverages a customer knowledge base and an innovative media resource function to deliver automatic voice services at Telecom Italia. The second presentation describes how virtual agents generate on-thefly VXML and grammars to support specific tasks at INPS (Italian Social Security). The third presentation discusses how to provide advanced natural language understanding features, such as implicit confirmation/correction, natural output of dynamic data, and calleradaptive pronunciation.

Help us celebrate the opening of SpeechTEK Europe and meet your speech colleagues from around the world at our informal drinks reception.